Integrated system for dynamic knowledge-based innovation using ai, knowledge graphs, and triz

Abstract

An integrated system for generating innovative solutions is disclosed. The system comprises a literature review AI agent, a knowledge graph, and a TRIZ Module. The AI agent continuously analyzes diverse information sources, feeding data into the knowledge graph. The knowledge graph organizes and connects information, identifying patterns and relationships. The TRIZ Module interacts with both components to identify contradictions and generate innovative solutions. This system accelerates innovation by processing vast amounts of information, recognizing cross-disciplinary patterns, and applying systematic problem-solving techniques.

Description

Attorney Docket No. 68036.2WO01INTEGRATED SYSTEM FOR DYNAMIC KNOWLEDGE-BASED INNOVATION USING Al, KNOWLEDGE GRAPHS, AND TRIZRELATED APPLICATIONS

[0001] The present application claims the benefit of U.S. Provisional Application No. 63 / 685,726, filed August 22, 2024, which is incorporated herein by reference in its entirety.FIELD OF THE DISCLOSURE

[0002] The disclosed technology relates to the field of artificial intelligence and innovation management, and more particularly to systems and methods for automating and enhancing the process of generating innovative solutions to complex problems across various domains.BACKGROUND

[0003] Innovation is crucial for technological advancement and solving complex global challenges. Traditional methods of innovation often rely on human expertise and are limited by the ability to process vast amounts of information. While artificial intelligence has been applied to various aspects of research and development, there remains a need for an integrated system that can continuously analyze information, identify patterns, and systematically generate innovative solutions.

[0004] Existing systems often lack the ability to dynamically update their knowledge base, struggle with cross-domain insights, and fail to apply systematic innovation principles effectively. The present disclosure addresses these limitations by integrating advanced Al, dynamic knowledge representation, and established innovation methodologies.SUMMARY

[0005] The present disclosure provides an integrated system for dynamic knowledge-based innovation that represents a paradigm shift in how complex problems are approached and solved across multiple domains. In accordance with one or more embodiments, the integrated system includes three main components that work synergistically to create a comprehensive innovation ecosystem. The firstAttorney Docket No. 68036.2WO01 component is a literature review artificial intelligent (Al) agent that may be configured to continuously scan and analyze diverse information sources, employing advanced natural language processing and machine learning algorithms to extract meaningful insights from scientific literature, patents, technical publications, and other knowledge repositories in real-time. The second component is a dynamic knowledge graph that may be configured to organize and connect information using a sophisticated hypergraph data structure, identifying patterns and relationships across disparate domains while continuously evolving its structure to accommodate new knowledge and generate novel hypotheses through semantic reasoning. The third component is a theory of inventive problem solving (TRIZ) module that may be configured to interact with the other components to systematically identify technical and physical contradictions within the knowledge domain and generate innovative solutions by applying the 40 inventive principles of TRIZ methodology, enhanced by cross-domain analogical reasoning and machine learning optimization techniques.

[0006] In various embodiments, the integrated system operates through a continuous feedback loop where each component informs and enhances the others, creating an emergent intelligence that exceeds the sum of its parts. The literature review Al agent feeds processed information into the dynamic knowledge graph, which then provides contextual knowledge to the TRIZ module for solution generation. Conversely, solutions and insights generated by the TRIZ module are integrated back into the knowledge graph, enriching the system's understanding and capability over time.

[0007] The system's operation is characterized by several key processes: continuously updating its knowledge base through automated information acquisition and processing, recognizing cross-disciplinary patterns that might be invisible to human researchers constrained by specialization boundaries, identifying problems and contradictions through advanced analytics and semantic analysis, and generating solutions using TRIZ principles augmented by the vast interconnected knowledge repository. The system's ability to operate across multiple domains simultaneously enables it to identify non-obvious connections and transfer solutions from one field to another, potentially leading to breakthrough innovations.

[0008] It is designed to work collaboratively with human inventors through an intuitive human-AI collaboration interface that includes interactive visualizationAttorney Docket No. 68036.2WO01 tools, natural language query capabilities, and explanation generation mechanisms. This collaborative approach ensures that the system augments rather than replaces human creativity, presenting findings and suggestions in an accessible manner while incorporating human feedback to continuously improve its performance and relevance. The system maintains transparency in its reasoning processes, allowing users to understand how conclusions are reached and enabling them to guide the innovation process based on their domain expertise and practical constraints.

[0009] Additional aspects, features, and advantages of the present disclosure will become apparent from the following detailed description.BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The accompanying drawing figures incorporated in and forming a part of this specification illustrate several aspects of the disclosure, and together with the description, serve to explain the principles of the disclosure.

[0011] FIG. 1 illustrates a block diagram of an integrated system architecture, in accordance with embodiments of the present disclosure.

[0012] FIG. 2 shows process steps the literature review Al agent collects, processes, analyzes, and extracts knowledge from various data sources, culminating in the output of structured information to the dynamic knowledge graph, in accordance with embodiments of the present disclosure.

[0013] FIG. 3 shows a cross-disciplinary knowledge graph, which visually represents the interconnections between various scientific disciplines, concepts, and methodologies, in accordance with embodiments of the present disclosure.

[0014] FIGs. 4A and 4B show how concept nodes is connected to disciplines and sub-disciplines in the knowledge graph, in accordance with embodiments of the present disclosure.

[0015] FIG. 5 shows process steps for how the TRIZ module leverages information from the knowledge graph, applies TRIZ principles to identify and resolve contradictions, and generates innovative solutions, in accordance with embodiments of the present disclosure.

[0016] FIG. 6 illustrates a flow chart for a method, in accordance with embodiments of the present disclosure.

[0017] FIG. 7 illustrates a flow chart for a method, in accordance with embodiments of the present disclosure.Attorney Docket No. 68036.2WO01

[0018] FIG. 8 is a simplified diagram illustrating a neural network structure, according to some embodiments.

[0019] FIG. 9 is a block diagram of a computer system suitable for implementing one or more components or operations in FIGS. 1-8, according to some embodiments.DETAILED DESCRIPTION

[0020] For purposes of promoting an understanding of the principles of the present disclosure, reference will now be made to the embodiments illustrated in the drawings, and specific language will be used to describe the same. It is nevertheless understood that no limitation to the scope of the disclosure is intended. Any alterations and further modifications to the described systems, devices, and methods, and any further application of the principles of the present disclosure are fully contemplated and included within the present disclosure as would normally occur to one skilled in the art to which the disclosure relates. In particular, it is fully contemplated that the features, components, and / or steps described with respect to one embodiment may be combined with the features, components, and / or steps described with respect to other embodiments of the present disclosure. For the sake of brevity, however, the numerous iterations of these combinations will not be described separately.

[0021] Embodiments of the present disclosure offer an advanced artificial intelligence system designed to revolutionize problem-solving and invention processes. By integrating knowledge graphs. TRIZ principles, and Al algorithms, the disclosed integrated system provides a powerful tool for identifying innovative solutions across diverse fields. This detailed description aims to provide a clear understanding of the invention's components and functionality. Various details of embodiments of the present disclosure are provided below, and additional benefits and / or other advantages will become apparent to those skilled in the art having benefit of the present disclosure.System Architecture

[0022] FIG. 1 illustrates a block diagram of an integrated system 10, which illustrates the system's structure, the flow of information between components, and how users and external systems interact with it, in accordance with embodiments of the present disclosure. As illustrated in FIG. 1, the system 10 combines threeAttorney Docket No. 68036.2WO01 powerful technologies, including literature review artificial intelligence (Al) agent 100, knowledge graph 101. and a theory of inventive problem solving (TRIZ) module 102.

[0023] In one or more implementations, the literature review Al agent 100 may be configured to continuously analyze diverse information sources, processing and extracting relevant data from scientific literature, patents, and other knowledge repositories. The literature review Al agent 100 may employ advanced natural language processing techniques and / or large language models, BERT models, and the like, to comprehend complex technical content across multiple domains. It can scan and process information from various sources including scientific journals, patent databases, conference proceedings, technical reports, social media feeds, news articles, and academic theses. The Al agent 100 may utilize machine learning algorithms to identify patterns, trends, and emerging developments that might not be apparent through manual review. In some aspects, the Al agent 100 can perform multi-lingual processing, enabling access to global knowledge sources. The literature review Al agent 100 may also incorporate sentiment analysis to gauge the significance of developments, named entity recognition to identify key actors and technologies, and relationship extraction to map connections between concepts. This component serves as the primary information gathering mechanism for the system, ensuring that the knowledge graph is populated with current and comprehensive data from across the scientific and technical landscape.

[0024] In one or more implementations, knowledge graph 101 (also referred to herein as the dynamic knowledge graph) may be configured to serve as the central repository, this component organizes and connects information, identifying patterns and relationships across various domains. The knowledge graph 101 may employ a hypergraph data structure that allows for multi-way relationships between nodes, enabling complex knowledge representation. In some aspects, the graph continuously evolves its structure to accommodate new information fed by the literature review Al agent 100, forming new connections and identifying emerging patterns through semantic reasoning. The knowledge graph may include nodes representing entities and concepts, edges representing relationships between nodes, and attributes providing additional metadata. In some cases, the graph utilizes advanced algorithms to identify cross-domain connections that might not be apparent to human researchers constrained by disciplinary boundaries. The graphAttorney Docket No. 68036.2WO01101's self-organizing nature allows it to adapt to changing knowledge landscapes, clustering related concepts hierarchically and employing dimensional reduction techniques for efficient storage and retrieval. The knowledge graph 101 may also incorporate ontology integration, RDF triple stores, and OWL reasoning capabilities to enhance its semantic understanding and inference capabilities.

[0025] In one or more implementations, the TRIZ Module 102 may be configured to connect via a network or another application programming interface and interact with both the Al Agent 100 and knowledge graph 101 to identify contradictions and generate innovative solutions based on TRIZ principles. The TRIZ Module 102 may employ advanced algorithms to automatically identify technical and physical contradictions within the knowledge domain. In some aspects, the module implements a neural network-based approach to apply the 40 TRIZ Inventive Principles to identified contradictions. The module may leverage the interconnected nature of the knowledge graph to suggest solutions from analogous problems in different domains, enabling cross-disciplinary innovation. In some cases, the TRIZ Module 102 utilizes genetic algorithms for solution optimization, reinforcement learning for strategy improvement, and collaborative filtering for solution ranking. The module maintains a dynamic connection with the knowledge graph 101 to ensure it operates with the most current information, including a real-time query interface and knowledge update listener. To facilitate interaction with human inventors and domain experts, the TRIZ module 102 may include an explanation generator with a user interface which receives and displays clear, human-readable explanations for the TRIZ module 102's reasoning and suggestions, an interactive solution explorer, and a feedback incorporation mechanism that integrates human feedback to guide and refine the problem-solving process.

[0026] These core components communicate with and are integrated via a user- friendly interface 103 that allows for input, comprehensive input, interactive visualization, and seamless interaction with the other components in system 10. The interface 103 may provide dynamic, customizable dashboards, natural language query capabilities, and intuitive controls for navigating the knowledge space accessible via gram 101 and TRIZ module 102. The architecture also incorporates robust connections to external data sources 104 and such as scientific databases, patent repositories, and real-time information feeds, as well as legacy systems 105,Attorney Docket No. 68036.2WO01 ensuring comprehensive data input and compatibility with existing organizational infrastructures and databases.

[0027] In a non-limiting embodiment, system 10 may be encapsulated within a secure framework 106, emphasizing the importance of data protection and integrity. The secure framework 106 may include a security layer that comprises encryption protocols, access control mechanisms, audit logging capabilities, and compliance features to safeguard sensitive information while allowing appropriate sharing and collaboration.

[0028] This integrated architecture enables the system to process vast amounts of information, recognize cross-disciplinary patterns, and apply systematic problemsolving techniques, thereby significantly accelerating the innovation process and potentially leading to breakthrough solutions in various fields. In some aspects, the system's components work synergistically, with each enhancing the capabilities of the others through continuous feedback loops and adaptive learning mechanisms. The modular design also allows for future expansion and integration of additional specialized components as new technologies and methodologies emerge.

[0029] FIG. 2 shows process steps of a method 20 of how the literature review Al agent 100 collects, processes, analyzes, and extracts knowledge from various data sources, culminating in the output of structured information to the dynamic knowledge graph, in accordance with embodiments of the present disclosure. As disclosed herein, the literature review Al agent 100 is an advanced artificial intelligence system designed to continuously scan and analyze diverse information sources. The agent employs natural language processing (NLP) and machine learning algorithms to extract key concepts, relationships, and emerging trends from the analyzed content. It is capable of understanding context, identifying significant developments, and recognizing subtle implications that might be overlooked by human researchers.

[0030] In various embodiments, the literature review Al agent 100 may include data sources 200, as shown in FIG. 2. In some implementations, the literature review Al agent 100 is capable of ingesting and processing information from a wide array of sources. These sources may include scientific articles from peer-reviewed journals, technical publications and white papers, patent databases, conference proceedings and presentations, and relevant social media discussions and forums. In some aspects, the Al agent 100 may also process industry reports and marketAttorney Docket No. 68036.2WO01 analyses, academic theses and dissertations, preprint servers (e.g., arXiv, bioRxiv), government research publications, and open-access databases. This diverse range of information sources allows the agent to maintain a comprehensive and up-to-date understanding of developments across multiple domains.

[0031] In some instances, Al agent 100 may crawl a network, such as the Internet through automated web scraping processes that systematically navigate and extract content from data sources 200, particularly those accessible via web pages. For example, Al agent 100 may access a list of target URLs or starting points, and use hyper-text transfer protocol (HTTP) requests to retrieve web page content. The Al agent 100 may parse the Hyper Text Markup Language (HTML) structures from which a web page is generated, and follow hyperlinks embedded in the HTML structures to discover additional pages, extract relevant text, images, or data elements based on predefined criteria. The Al agent 100 may also handle various web technologies such as JavaScript-rendered content, cookies, and authentication mechanisms.

[0032] During the content extraction phase, the Al agent 100 may employ natural language processing techniques to analyze and categorize the retrieved information. The Al agent 100 may filter content based on relevance, quality metrics, or specific domain requirements, all of which may be included in an Al agent schema or prompt. The Al agent 100 may then store the processed data in structured formats or incorporate the processed data into know graph 101 as will be discussed below.

[0033] As discussed above. Al agent 100 may identify new articles from data sources 200 that store data in a database or storage system. In some instances, Al agent 100 may use content fingerprinting techniques that generate unique hash values for each extracted article, enabling comparison against existing stored content to determine novelty. In another instances, the Al agent 100 may implement timestamp-based tracking mechanisms that monitor publication dates, last-modified headers, and content update indicators to identify recently published or modified articles. The Al agent 100 may utilize content similarity algorithms including cosine similarity calculations, Jaccard indices, and edit distance measurements to distinguish new articles from existing content variations or republished materials.Attorney Docket No. 68036.2WO01

[0034] In various embodiments, the literature review Al agent 100 may include Natural Language Processing (NLP) system 204, as shown in FIG. 2. In some implementations, the literature review Al agent 100 employs an NLP system 204 based on a transformer architecture that may include a GPT-based (Generative Pretrained Transformer) model, BERT model, or another NLP model. In some instances, the model may be fine-tuned for scientific and technical content understanding. The NLP system includes multi-lingual processing capabilities, domain-specific vocabulary understanding, and context-aware semantic analysis. It may perform sentiment analysis for identifying significant developments, named entity recognition for identifying key actors, organizations, and technologies, and relationship extraction for mapping connections between concepts.

[0035] In various embodiments, the literature review Al agent 100 may include machine learning algorithms 205, as shown in FIG. 2. In some implementations, the literature review Al agent 100 utilizes a combination of supervised and unsupervised machine learning algorithms to extract insights and identify patterns. These include: clustering algorithms (e.g.. K-means, DBSCAN) for grouping related concepts and identifying research themes, classification algorithms (e g.. Random Forests, Support Vector Machines) for categorizing content into predefined topics, anomaly detection algorithms to identify outlier information that may represent novel discoveries or emerging trends, and reinforcement learning algorithms for optimizing the agent's information gathering and analysis strategies over time.

[0036] In various embodiments, the literature review Al agent 100 may employ time series analysis and forecasting techniques to identify trends and patterns in the analyzed content, referred to herein as Trend and Pattern Recognition. This includes ARIMA (AutoRegressive Integrated Moving Average) models for trend analysis. Prophet algorithm for robust time series forecasting, particularly useful for identifying seasonal trends in research topics, and Latent Dirichlet Allocation (LDA) for topic modeling and tracking the evolution of research themes over time.

[0037] In various embodiments, the literature review Al agent 100 is designed with a self-improving architecture that allows it to enhance its performance over time, referred to herein as continuous learning and self-improvement. This is achieved through active learning techniques that allow the agent to identify areas where it needs additional training or human input, a feedback loop mechanism that incorporates corrections and validations from the knowledge graph and humanAttorney Docket No. 68036.2WO01 experts, and periodic retraining of the underlying models using transfer learning techniques to adapt to new domains and evolving language patterns.

[0038] In various embodiments, the literature review Al agent 100 also synthesizes information across multiple data sources, for example, for information extraction and synthesis. This is accomplished through cross-referencing algorithms that identify corroborating or conflicting information across multiple sources, abstractive summarization techniques that generate concise, coherent summaries of key findings and developments, and fact extraction and verification methods to ensure the accuracy of the information being fed into the knowledge graph.

[0039] In various embodiments, the literature review Al agent 100 is integrated with knowledge graph 101. In addition to feeding information into the knowledge graph 101, Al agent 100 also uses the existing knowledge structure to inform its analysis of data extracted from data sources 200. This bidirectional flow allows for: context-aware information processing, where new data is interpreted in light of existing knowledge, identification of knowledge gaps that can guide future information gathering priorities, and validation of new information against the established knowledge base.

[0040] To ensure the integrity and fairness of the information processing, the agent incorporates bias detection algorithms to identify and flag potential biases in the source material or in the agent's own processing, diverse source selection mechanisms to ensure a balanced representation of perspectives and origins, and transparency logs that record the provenance of information and the reasoning behind extracted insights.

[0041] In various implementations, FIG. 2 shows a flow chart for the process steps of the method 20 of the literature review Al agent 100. The process begins at the top of FIG. 2 with Al agent 100 accessing data sources 200. Data sources 200 may contain various information sources, including scientific journals, patent databases, conference proceedings, technical reports, social media feeds, news articles, and academic theses, and may include data in different formats, including text, images, and video.

[0042] From data sources 200, the process flows into a data ingestion operation 201, which has sub-processes, such as web scraping, API Integration, file parsing, and data cleaning, illustrating the methods used to collect and prepare the raw data.Attorney Docket No. 68036.2WO01

[0043] The data ingestion process 201 then connects to initial quality check decision 202. This has two outputs: a "Pass" arrow leading to the next process, and a "Fail" arrow leading to a "discard / flag for review" 203. This represents the initial filtering of data based on quality criteria.

[0044] The “Pass” leads to NLP processing 204, which comprises key natural language processing tasks, such as tokenization. named entity recognition, sentiment analysis, topic modeling, and semantic parsing.

[0045] The NLP Processing box is then connected to machine learning analysis process 205 which incorporates lists various machine learning algorithms 205 and machine learning techniques employed, such as classification, clustering, regression analysis, and anomaly detection. Classification may categorize extracted data into predefined classes or categories based on learned patterns and features. Clustering may group similar web content together without using predefined categories, discovering natural patterns and relationships within the extracted data. Regression analysis may predict continuous numerical values related to data characteristics, such as content quality scores, engagement metrics, or relevance rankings. Anomaly detection may identify unusual or suspicious patterns in the data that deviate from normal data, enabling the detection of fake data, incomplete data, and the like.

[0046] Following the machine learning analysis process 205 is knowledge extraction 206, which incorporates tasks like key concept identification, relationship mapping, and hypothesis formulation of the extracted data.

[0047] The final step in the flow chart is represented as an output to dynamic knowledge graph process 207. In this process, the data extracted in knowledge extraction 206 is incorporated into knowledge graph 101.

[0048] This flowchart illustrates the process of how the literature review Al agent 100 collects, processes, analyzes, and extracts knowdedge from various data sources, culminating in the output of structured information to the knowledge graph 101. It represents the data processing pipeline that may repeat continuously or at predefined intervals and enables the Al agent 100 to continuously learn and improve its performance in gathering and analyzing relevant information for the innovation system.

[0049] As disclosed herein, the literature review Al agent 100 serves as the primary input mechanism for the integrated innovation system. Its continuous operation ensures that the knowledge graph is always populated with the mostAttorney Docket No. 68036.2WO01 current and relevant information across various domains. This up-to-date and comprehensive knowledge base is then leveraged by the TRIZ Module to identify contradictions and generate innovative solutions.

[0050] In one or more embodiments, the knowledge graph 101 serves as the central repository of information for the system. Unlike traditional databases, this knowledge graph is a complex, interconnected network of concepts, entities, and relationships. As the literature review Al agent 100 feeds new information into the graph, it continuously updates and reorganizes itself. The graph uses advanced semantic reasoning to form new connections, identify patterns, and generate hypotheses.

[0051] In one or more embodiments, the knowledge graph 101 is implemented using a hypergraph data structure, allowing for the representation of multi-way relationships. It includes nodes representing entities, concepts, or data points; edges representing relationships between nodes; hyperedges representing complex relationships involving multiple nodes; and attributes providing additional metadata associated with nodes and edges.

[0052] The graph employs a dynamic structure that continuously evolves based on incoming information. This includes adaptive node creation where new nodes are automatically created for novel concepts or entities; dynamic edge weighting where relationship strengths are continuously updated based on frequency and recency of connections; hierarchical clustering where nodes are organized into hierarchical clusters to represent different levels of abstraction; and dimensional reduction where high-dimensional data is mapped to lower-dimensional representations for efficient storage and retrieval.

[0053] The graph maintains consistency and relevance through real-time updates. This may be accomplished through an incremental update algorithm that allows for efficient integration of new information without full graph recomputation; a conflict resolution mechanism that resolves contradictions in incoming data through probabilistic reasoning: temporal versioning that maintains historical versions of the graph to track knowledge evolution; and a distributed update protocol that enables parallel processing of updates across multiple computational nodes.

[0054] The graph incorporates semantic technologies to derive meaning and inferences. These may include ontology integration that incorporates domainspecific ontologies to provide semantic context; RDF (Resource DescriptionAttorney Docket No. 68036.2WO01Framework) triple store that represents knowledge in subject-predicate-object format; SPARQL query engine that enables complex semantic queries across the graph; and OWL (Web Ontology Language) reasoning that applies description logic to infer implicit knowledge.

[0055] The graph can identify non-obvious connections across different domains through similarity metrics that compute semantic and structural similarities between nodes across domains; path analysis that identifies meaningful paths between seemingly unrelated concepts; analogy detection that recognizes structural analogies between different domain subgraphs; and fusion algorithms that combine information from multiple domains to create cross-disciplinary insights.

[0056] The graph may actively generate hypotheses based on its structure and content through pattern recognition that identifies recurring patterns in subgraph structures; anomaly detection that flags unusual or missing connections that might indicate new hypotheses; predictive modeling that uses graph neural networks to predict potential new nodes or edges; and abductive reasoning that generates explanatory hypotheses for observed phenomena in the graph.

[0057] To handle vast amounts of data and complex operations, the graph may implement distributed graph processing that utilizes frameworks like Apache Graph for large-scale graph computations; in-memory graph database that employs high- performance in-memory storage for frequently accessed subgraphs; adaptive indexing that dynamically creates and maintains indices based on query patterns; and approximate query processing that provides fast, approximate results for complex queries when exact results are not required.

[0058] The graph is tightly integrated with other Al components of the system through a bidirectional API that allows for efficient data exchange with the literature review Al agent 100; a query interface that provides a sophisticated query language for the TRIZ Module to extract relevant information; and a feedback loop that incorporates insights and solutions generated by the TRIZ Module back into the graph structure.

[0059] To facilitate human-AI collaboration, the graph may include interactive visualization tools that allow users to explore and navigate the graph structure; a natural language interface that enables querying and interaction with the graph using natural language; and explanation generation that provides human-readable explanations for inferences and hypotheses.Attorney Docket No. 68036.2WO01

[0060] To protect sensitive information and ensure data integrity, the graph may implement granular access control that allows fine-grained permissions on nodes and subgraphs; encryption that applies encryption to sensitive data both at rest and in transit; and audit trail that maintains a secure log of all modifications and accesses to the graph.

[0061] As disclosed herein, the knowledge graph 101 serves as the central nervous system of the integrated innovation system. It not only stores and organizes the vast amount of information collected by the literature review Al agent 100 but also actively processes this information to generate new insights and hypotheses.

[0062] FIG. 3 shows an example of a cross-disciplinary knowledge graph 30, which visually represents the interconnections between various scientific disciplines, concepts, and methodologies. It illustrates the potential for identifying novel connections between disciplines, highlighting emerging research areas, and facilitating interdisciplinary insights that could lead to scientific breakthroughs. It contains a central hub the interdisciplinary nexus 330 which serves as the focal point, major discipline nodes including Physics 331, Chemistry 332, Biology 333, Computer Science 334, Mathematics 335, Environmental Science 336, Neuroscience 337, and Materials Science 338, sub-discipline nodes for example Quantum Mechanics 339 under Physics, Organic Chemistry 340 under Chemistry, Genetics 341 under Biology, and Machine Learning 342 under Computer Science, and concept nodes connected to disciplines and sub-disciplines for example in FIG. 4A the concept node Quantum Entanglement 347 is connected to the discipline Computer Science 334 and its sub-discipline Quantum Computing 243. and it is also connected to the discipline Physics 331 and its sub-discipline Quantum Mechanics 339. FIGs. 4A and 4B are diagrams illustrating how concept nodes are connected to disciplines and sub-disciplines in the knowledge graph, in accordance with embodiments of the present disclosure. Similarly, in FIG. 4B concept node Protein Folding 348 is connected to the discipline Biology 333 and its sub-discipline Molecular Biology 346, and is connected to the discipline Chemistry 332 and its sub-discipline Biochemistry 345, it is also connected to the discipline Computer Science 334 and its sub-discipline bioinformatics 344.

[0063] The graph 30's ability to identify cross-domain connections and generate hypotheses is particularly crucial for the innovation process. By recognizing non- obvious relationships between concepts from different fields, it can suggest novelAttorney Docket No. 68036.2WO01 approaches to problems that might not be apparent to human researchers constrained by disciplinary boundaries.

[0064] Furthermore, the graph 30's dynamic nature ensures that it always represents the current state of knowledge across various domains. This up-to-date, interconnected knowledge base provides the TRIZ Module with rich, contextual information to identify contradictions and generate innovative solutions.

[0065] The self-organizing and self-improving nature of the graph also means that its value increases over time. As more information is added and processed, the graph becomes increasingly adept at recognizing patterns, generating insights, and supporting the innovation process.

[0066] Now referring to FIG. 5, which shows steps for the process 40 of how a Theory of Inventive Problem Solving (TRIZ) module leverages information from the knowledge graph, applies TRIZ principles to identify and resolve contradictions, and generates innovative solutions, in accordance with embodiments of the present disclosure.

[0067] As disclosed herein, the TRIZ Module is an implementation of the systematic innovation methodology, enhanced by its integration with the other components of the system. This module interacts with both the literature review Al agent 100 and the knowledge graph 101 to identify contradictions, apply inventive principles, and propose innovative solutions, as described with respect to FIG. 1.

[0068] In some implementations, the TRIZ Module 102 incorporates several sophisticated components to identify contradictions and generate innovative solutions. For automated contradiction identification, the module may employ advanced algorithms to detect both technical and physical contradictions within the knowledge domain. These algorithms may include semantic analysis that utilizes natural language processing to identify opposing requirements in textual descriptions, with parameters such as linguistic patterns, domain-specific terminology, and contextual relevance scores. Graph-based contradiction detection may analyze the knowledge graph structure to identify nodes with conflicting attributes or relationships, using parameters like node attribute thresholds, edge weight differentials, and subgraph pattern matching criteria. Time series contradiction identification can detect contradictions in temporal data, such as conflicting trends or cyclical patterns, with parameters including time window size, trend reversal thresholds, and seasonality factors.Attorney Docket No. 68036.2WO01

[0069] For applying TRIZ principles, the module may implement a neural network-based approach to leverage the 40 TRIZ Inventive Principles. This approach can include a principle mapping neural network, which is a deep learning model trained on historical inventions and patents to map problem characteristics to relevant TRIZ principles, with parameters such as input feature vector (problem characteristics), hidden layer architecture, activation functions, and output layer (principle relevance scores). Contextual principle adaptation may adjust the application of principles based on the specific domain and context of the problem, using parameters such as domain weightings, contextual relevance factors, and principle combination probabilities. A principle execution simulator can simulate the application of selected principles to predict potential outcomes and side effects, with parameters including principle application rules, system model parameters, and simulation time steps.

[0070] The TRIZ module 102 leverages the interconnected nature of the knowledge graph to suggest solutions from analogous problems in different domains. The module 102 may include an analogical reasoning engine that identifies structurally similar problems across different domains in the knowledge graph, utilizing parameters such as graph similarity metrics, domain distance measures, and analogy strength thresholds. In some aspects, the module incorporates a Solution Transfer Algorithm that adapts solutions from analogous problems to fit the current problem context, with parameters including solution abstraction level, domain-specific constraints, and adaptation rules. The module may also feature a Novelty Scoring Mechanism that evaluates the novelty of suggested solutions based on their distance from existing solutions in the problem domain, employing parameters such as novelty threshold, domain familiarity index, and solution feature vector comparisons. Through these components, the module can effectively identify and adapt solutions from one field to address challenges in another, potentially enabling cross-disciplinary innovation.

[0071] The TRIZ module 102 maintains a dynamic connection with the knowledge graph to ensure it operates on the most current information. This connection may include a Real-time Query Interface that allows the TRIZ module 102 to perform complex queries on the knowledge graph during the problemsolving process, with parameters such as query complexity limits, response time thresholds, and relevance scoring criteria. In some aspects, the TRIZ module 102Attorney Docket No. 68036.2WO01 incorporates a Knowledge Update Listener that continuously monitors for relevant updates in the knowledge graph that might affect ongoing problem-solving processes, utilizing parameters including update relevance thresholds, notification frequency, and impact assessment criteria. The TRIZ module 102 may also feature a Dynamic Constraint Updater that adjusts problem constraints and parameters based on the latest information from the knowledge graph, with parameters such as constraint update frequency, confidence thresholds for new information, and constraint conflict resolution rules.

[0072] The TRIZ module 102 employs an iterative approach to continuously improve and refine proposed solutions. This approach may include a Genetic Algorithm for Solution Optimization that evolves and refines solution candidates over multiple generations, utilizing parameters such as population size, mutation rate, crossover probability, and fitness function definition. In some aspects, the TRIZ module 102 incorporates Reinforcement Learning for Strategy Improvement, which learns and adapts problem-solving strategies based on the success of previous solutions, with parameters including learning rate, discount factor, and exploration versus exploitation balance. Additionally, the TRIZ module 102 may implement Collaborative Filtering for Solution Ranking that ranks and filters solution candidates based on their predicted effectiveness and feasibility, employing parameters such as similarity metrics, rating scale, and neighbor selection criteria. Through these iterative refinement mechanisms, the module can progressively enhance the quality and applicability of the generated innovative solutions over time.

[0073] To facilitate interaction with user, including inventors and domain experts, the module may include an Explanation Generator that translates the output of TRIZ module 102, including reasoning and suggestions into clear, human- readable explanations, with parameters such as explanation detail level, user expertise profile, and domain-specific terminology usage. In some aspects, the TRIZ module 102 incorporates an Interactive Solution Explorer that allows users to navigate and manipulate proposed solutions in a visual interface, utilizing parameters including visualization complexity, interaction depth, and real-time update frequency. Additionally, the module may feature a Feedback Incorporation Mechanism that integrates human feedback to guide and refine the module'sAttorney Docket No. 68036.2WO01 problem-solving process, employing parameters such as feedback weight factors, consistency checking thresholds, and learning rate from human input.

[0074] Further details of the process 40 of how TRIZ module leverages information from the knowledge graph, applies TRIZ principles to identify and resolve contradictions, and generates innovative solutions.

[0075] As illustrated in FIG. 5. the process 40 begins with a Problem Input from knowledge graph 400 which includes problem statement, context, and constraints in a human language.

[0076] The Problem Input from knowledge graph 400 serves as the entry point for the TRIZ analysis process, receiving structured problem statements identified by the knowledge graph 101. This component processes cross-disciplinary patterns detected in the knowledge graph, user-submitted innovation challenges, and automatically identified research gaps or technological bottlenecks. For example, the system might receive a problem statement such as "Increase the efficiency of solar panel energy conversion while reducing manufacturing costs and maintaining durability in harsh weather conditions." The component contextualizes this problem within the renewable energy industry and manufacturing scalability requirements, while identifying constraints such as material cost limits, existing manufacturing processes, and regulatory standards. Performance targets might include achieving greater than 22% efficiency and manufacturing costs below $0.50 per watt. The output from Problem Input from knowledge graph (400) is a structured problem definition with clearly identified objectives, constraints, and success criteria formatted specifically for TRIZ analysis.

[0077] The next step is Problem Analysis 401 which includes parameter identification, system analysis, and functional analysis. Problem Analysis 401 receives the structured problem statement from Problem Input from knowledge graph 400 and systematically breaks down the problem into analyzable components using established TRIZ methodology. The parameter identification process maps problem elements to TRIZ's 39 standard engineering parameters, which are measurable characteristics that describe the system such as weight, speed, strength, cost, and efficiency.

[0078] For the solar panel example, these parameters would include conversion efficiency (percentage), cost per watt (dollars), weight (kg / m2), and durability (years). The system analysis component identifies the super-system (entireAttorney Docket No. 68036.2WO01 renewable energy infrastructure), the system itself (solar panel assembly), and subsystems (photovoltaic cells, protective coating, frame), establishing system boundaries and hierarchical relationships. Functional analysis examines useful functions (what the system should do, such as converting sunlight to electricity), harmful functions (undesired effects like heat generation and degradation), and insufficient functions (underperforming aspects such as low light conversion). This process creates function models that establish cause-and-effect relationships between system elements. The output from Problem Analysis (401) is a comprehensive system model with identified parameters, functions, and system hierarchy.

[0079] The next step is Contradiction Identification 402 which includes a decision that separates the process into two types of contradictions: technical contradictions and physical contradictions. Contradiction Identification 402 analyzes the problem model from Problem Analysis 401 to identify specific types of contradictions using TRIZ classification systems. The component distinguishes between technical contradictions, where improving one parameter leads to degradation of another, and physical contradictions, where a system requires opposite states of the same parameter. For technical contradictions, the identification criteria focus on situations when enhancing Parameter A orsens Parameter B. such as increasing solar panel efficiency (Parameter A) while increasing manufacturing complexity and cost (Parameter B). Physical contradictions are identified when the same object needs contradictory properties, such as a solar panel surface needing to be transparent for light penetration AND opaque for maximum photon absorption. The decision-making algorithm analyzes parameter relationships, checks for parameter conflicts versus state conflicts, applies TRIZ contradiction matrix lookup procedures, and prioritizes contradictions by their impact on system performance. The output from Contradiction Identification (402) consists of classified contradictions with severity rankings and associated TRIZ parameters.

[0080] For technical contradictions, the process includes a representation of the Contradiction Matrix 403. Contradiction Matrix 403 maps the identified contradictions from Contradiction Identification 402 to TRIZ inventive principles using the classical TRIZ contradiction matrix. The process involves matrix lookup procedures that cross-reference improving versus worsening parameters, principleAttorney Docket No. 68036.2WO01 selection that retrieves recommended inventive principles (typically 2-4 principles per contradiction), and relevance scoring that ranks principles based on knowledge graph insights and domain-specific success rates. For the solar panel example, a contradiction between efficiency and cost would be positioned in the matrix as Parameter 22 (Energy use by stationary object) versus Parameter 32 (Ease of manufacture), suggesting principles such as # 1 (Segmentation), #15 (Dynamics), and #40 (Composite materials). The output from Contradiction Matrix 403 is a prioritized list of applicable inventive principles with relevance scores.

[0081] Next is the 40 Inventive Principles 404 of TRIZ which would highlight the mechanism to show how specific principles are selected based on the identified contradiction. Inventive Principles 404 provides detailed implementation guidance for the selected TRIZ principles from Contradiction Matrix 403. For each principle, the component provides principle definitions with core concept explanations, subprinciples offering specific implementation approaches, domain applications showing how the principle applies to the current problem context, and historical examples of successful implementations drawn from the knowledge graph. For example, Principle #40 (Composite Materials) would include guidance to use composite or layered structures, apply materials with complementary properties, implement multi-junction solar cells with different band gaps, and expect benefits of higher efficiency across broader light spectrum. The output from Inventive Principles 404 consists of actionable principle descriptions with comprehensive implementation strategies.

[0082] Next step in the process is the Principle Application 405 which translates the abstract TRIZ principles from Inventive Principles 404 into concrete solution concepts through contextualization that adapts principles to specific problem domains, combination logic that merges multiple principles for synergistic effects, feasibility filtering that screens concepts against technical and economic constraints, and concept development that creates detailed solution architectures. An example application might combine Segmentation and Composite Materials principles to create a modular solar panel design with different cell types optimized for different light conditions, implementing morning / evening panels optimized for low-light conditions alongside midday panels optimized for high-intensity conditions. The output from Principle Application 405 includes multiple solution concepts with detailed technical specifications.Attorney Docket No. 68036.2WO01

[0083] This leads to Solution Generation 406, which includes idea formulation, concept development, and solution modeling. Solution Generation 406 creates comprehensive solution packages from the applied principles developed in Principle Application 405. The component produces technical specifications with detailed design parameters, implementation roadmaps with step-by-step development processes, resource requirements including materials, manufacturing processes, and expertise needed, and performance predictions with expected improvements in target parameters. An example solution might be an "Adaptive Multi-Layer Solar Panel System: Three-layer photovoltaic design with dynamic light-routing mechanisms, incorporating perovskite top layer for UV spectrum, silicon middle layer for visible light, and gallium arsenide bottom layer for IR spectrum." The output from Solution Generation 406 consists of complete solution descriptions with comprehensive implementation details.

[0084] The next step is Solution Evaluation 407, which includes criteria such as novelty, feasibility, and effectiveness. Solution Evaluation 407 assesses the generated solutions from Solution Generation 406 using multiple criteria including novelty assessment through patent landscape analysis via knowledge graph, prior art comparison, and innovation scoring based on concept uniqueness. Feasibility analysis includes technical readiness level evaluation, manufacturing capability assessment, regulatory compliance checks, and market adoption barrier analysis. Effectiveness measurement quantifies performance improvements, conducts costbenefit analysis, performs risk assessment, and evaluates scalability potential. The scoring method employs multi-criteria decision analysis with weighted scoring across all dimensions. The output from Solution Evaluation 407 includes ranked solutions with detailed evaluation reports and recommendation confidence levels.

[0085] This leads to Solutions Refinement Loop 408. Solution Refinement Loop 408 iteratively improves solutions based on evaluation feedback from Solution Evaluation 407 through gap analysis that identifies weaknesses in current solutions, parameter adjustment that modifies technical specifications, alternative exploration that generates solution variants, re-evaluation that tests refined concepts against criteria, and convergence checking that determines if further refinement is beneficial. The refinement process continues until termination criteria are met, including improvement rate falling below threshold, maximum iteration count reached, feasibility constraints satisfied, or stakeholder approval obtained. TheAttorney Docket No. 68036.2WO01 output from Solution Refinement Loop 408 consists of optimized solution candidates ready for final presentation.

[0086] The final element of the process is an output which is the Proposed Innovative Solutions 409. Proposed Innovative Solutions 409 presents the final, validated innovation recommendations from Solution Refinement Loop 408 in comprehensive solution packages. Each package includes an executive summary’ with problem statement recap, solution overview and key innovations, and expected benefits and success metrics. Technical specifications provide detailed design parameters, performance characteristics, and material and component requirements. The implementation plan outlines development timeline with milestones, resource allocation requirements, risk mitigation strategies, and validation and testing protocols. The business case includes market opportunity analysis, competitive advantage assessment, investment requirements and ROI projections, and intellectual property strategy. For example, a solar panel innovation might be titled "Spectrum- Adaptive Modular Solar System" featuring dynamic three-layer photovoltaic architecture with intelligent light routing, offering 35% efficiency improvement over conventional panels, 20% reduction in manufacturing cost through modular design, and extended 30-year lifespan through adaptive materials. The implementation roadmap would include Phase 1 prototype development and testing (6 months), Phase 2 pilot manufacturing setup (12 months), and Phase 3 market validation and scaling (18 months), with expected high market disruption potential, 85% technical feasibility confidence, and strong patent strength with 12 novel claims identified. The output from Proposed Innovative Solutions 409 is a comprehensive innovation dossier ready for decision-making, patent filing, or development initiation.

[0087] The process 40 illustrated in FIG. 5 shows how the TRIZ module 102 leverages information from the knowledge graph, applies TRIZ principles to identify and resolve contradictions, and generates innovative solutions. As described herein, the TRIZ module 102 serves as the creative engine of the integrated innovation system. It leverages the vast and up-to-date knowledge provided by the literature review Al agent 100 and the knowledge graph 101 to systematically generate innovative solutions to complex problems. The TRIZ module 102's ability to automatically identify contradictions across diverse domains allows it to tackle problems that might not be apparent to human researchers limitedAttorney Docket No. 68036.2WO01 by specialization. By applying TRIZ principles in conjunction with cross-domain knowledge, the module can suggest truly novel solutions that bridge different fields of knowledge.

[0088] The iterative refinement process, combined with the ability to incorporate human feedback, ensures that the generated solutions are not only innovative but also practical and implementable. This creates a powerful synergy between artificial intelligence and human creativity, potentially accelerating the pace of innovation across multiple industries.

[0089] Furthermore, the module's integration with the constantly updating knowledge graph means that it can adapt its problem-solving strategies in real-time as new information becomes available. This ensures that the innovation process is always informed by the latest developments and discoveries across relevant fields.

[0090] In accordance with one or more embodiments, the disclosed integrated system, such as, system 10 as described with respect to FIG. 1, for dynamic knowledge-based innovation operates through a series of interconnected processes, each leveraging the capabilities of its core components to drive continuous innovation. The system's operation is described in further details as follows.

[0091] In various embodiments, the system 10 may maintain an up-to-date knowledge base through constant information acquisition and processing. In some embodiments, the literature review Al agent 100 may perform regular, scheduled scans of predefined information sources at configurable intervals, with parameters including scan frequency, source priority levels, and data volume thresholds. In some aspects, continuous monitoring of high-priority sources enables immediate integration of breaking developments, utilizing parameters such as update latency, relevance thresholds, and source credibility scores. New information can be seamlessly incorporated into the knowledge graph using delta update algorithms, with parameters that may include update batch size, conflict resolution rules, and integration confidence thresholds. In some cases, the system adjusts its learning rate based on the volume and significance of new information, employing parameters such as learning rate bounds, significance scoring criteria, and domain-specific adjustment factors.

[0092] The knowledge graph 101 employs advanced analytics to identify emerging patterns and trends. The system may utilize temporal pattern analysis through time series analysis algorithms that detect trends, cycles, and anomalies inAttorney Docket No. 68036.2WO01 the evolution of concepts and relationships, with parameters including time window sizes, trend significance thresholds, and seasonality detection criteria. In some aspects, topological pattern mining is implemented through graph analytics algorithms that identify recurring subgraph structures indicative of important patterns, utilizing parameters such as subgraph size limits, occurrence frequency thresholds, and isomorphism criteria. The system may also perform cross-domain correlation analysis using statistical methods to assess correlations between developments in different domains, with parameters including correlation strength thresholds, lag period ranges, and false discovery rate control. Additionally, in some cases, emergent property detection is conducted through complexity analysis techniques that identify emergent properties arising from the interaction of multiple factors, employing parameters such as complexity metrics, emergence criteria, and system boundary definitions.

[0093] The system proactively identifies areas requiring innovation: through several complementary approaches. A Gap Analysis Algorithm may compare the current state of knowledge against projected advancement trajectories to identify knowledge gaps, utilizing parameters such as projection time horizons, confidence intervals, and domain-specific growth models. In some aspects, Contradiction Detection performs automated analysis of the knowledge graph structure to identify logical or physical contradictions, with parameters including contradiction types, severity scoring, and context consideration factors. The system may also incorporate Opportunity Scoring that evaluates identified problems based on their potential impact and feasibility of solution, employing impact metrics, feasibility criteria, and domain-specific weighting factors. Additionally, in some cases, Novelty Assessment measures the uniqueness of identified problems against the existing body of knowledge, using parameters such as similarity thresholds, prior art search depth, and novelty scoring functions.

[0094] The TRIZ module 102 leverages identified problems and comprehensive knowledge to generate innovative solutions through several complementary approaches. The module may implement problem-principle mapping that matches identified problems to relevant TRIZ inventive principles using machine learning classifiers, with parameters including feature extraction rules, classification confidence thresholds, and multi-label classification settings. In some aspects, the module utilizes solution space exploration through genetic algorithms to exploreAttorney Docket No. 68036.2WO01 and optimize the solution space defined by applicable TRIZ principles, employing parameters such as population size, mutation rates, fitness function definitions, and convergence criteria. The TRIZ module 102 may also incorporate analogical solution transfer that adapts solutions from analogous problems in different domains identified in the knowledge graph, with parameters including analogy strength thresholds, transfer learning parameters, and domain adaptation rules. Additionally, in some cases, the module implements a constraint satisfaction solver to ensure generated solutions meet all identified constraints and requirements, utilizing parameters such as constraint prioritization, relaxation criteria, and solution feasibility thresholds.

[0095] The system continuously improves its solutions through iterative feedback and analysis. This may be accomplished through a Solution Evaluation Algorithm that assesses generated solutions based on novelty, effectiveness, and feasibility, utilizing parameters such as evaluation criteria weights, multi-objective optimization settings, and Pareto frontier analysis. In some aspects, a knowledge graph Update Protocol integrates promising solutions and their implications back into the knowledge graph, employing parameters including integration confidence thresholds, relationship strength initialization, and concept novelty assessment. The system may also implement a Recursive Improvement Process that applies the problem-solving process to its own outputs to identify and resolve secondary issues, with parameters such as recursion depth limits, improvement significance thresholds, and termination criteria. Additionally, in some cases, the system incorporates Learning from Feedback mechanisms that utilize success / failure data of implemented solutions to refine future problem-solving strategies, with parameters including feedback weighting factors, long-term vs. short-term learning balance, and forgetting rate for outdated information.

[0096] The system facilitates effective collaboration between Al and human inventors through several integrated components. An Intuitive Visualization Engine may generate interactive, multi-dimensional visualizations of the knowledge space, identified problems, and proposed solutions, with parameters including visualization complexity levels, user preference learning, and real-time interaction capabilities. The system may incorporate a Natural Language Interface that provides a conversational Al interface for users to query, guide, and interact with the system, utilizing parameters such as language understanding confidenceAttorney Docket No. 68036.2WO01 thresholds, context retention settings, and personality adaptation factors. In some aspects, an Explanation Generation Module produces clear, contextualized explanations for the system's reasoning and suggestions, with parameters including explanation detail levels, user expertise profiling, and domain-specific terminology usage. The system may also feature a Collaborative Editing Environment that allows human users to directly modify and refine Al-generated solutions, employing parameters such as version control settings, conflict resolution protocols, and attribution tracking. Additionally, in some cases, a Feedback Learning Mechanism continuously learns from human feedback to improve the relevance and quality of its outputs, utilizing parameters including feedback incorporation rates, consistency checking thresholds, and long-term preference modeling.

[0097] This operational framework enables the system to function as a powerful, self-improving engine for innovation. By continuously learning, identifying patterns, recognizing problems, generating solutions, and collaborating with human experts, the system can tackle complex, multidisciplinary challenges and drive innovation across various fields of knowledge.

[0098] The system's ability to operate across domains, identify non-obvious connections, and apply systematic innovation principles makes it a unique and powerful tool for accelerating technological advancement and solving pressing global challenges. Its iterative, self-improving nature ensures that its capabilities grow over time, potentially leading to exponential growth in innovation capacity.Examples

[0099] To demonstrate effectiveness of the disclosed technology, a real-world example of how this integrated system for dynamic knowledge-based innovation might be used to address a significant global challenge for example developing more efficient and sustainable urban transportation systems.

[0100] Initial User Input: A city planner inputs the following query into the system: "How can we improve urban transportation to reduce congestion and carbon emissions while enhancing mobility for all citizens?" Step 1 : Literature Review Al Agent Activation

[0101] The system's literature review7Al agent 100 begins by scanning and analyzing relevant sources: a. Urban planning journals b. Transportation engineering publicationsAttorney Docket No. 68036.2WO01 c. Environmental science reports d. Smart city technology whitepapers e. Public policy documents f. Social media discussions on urban mobility

[0102] The agent processes this information, extracting key concepts, trends, and emerging technologies related to urban transportation.Step 2: knowledge graph Update

[0103] The processed information is integrated into the knowledge graph 101, updating and expanding its representation of urban transportation systems. This includes: a. Current transportation modes and their efficiencies b. Environmental impact data c. Population density and movement patterns d. Emerging technologies like autonomous vehicles and smart traffic management systems e. Successful case studies from various cities worldwideStep 3: Pattern Recognition and Problem Identification

[0104] The knowledge graph identifies several key patterns and issues: a. Increasing urbanization is exacerbating traffic congestion b. Current public transportation systems are often inefficient and underutilized c. There's a growing trend towards micromobility solutions (e-scooters, bikesharing) d. Electric vehicle adoption is rising but faces infrastructure challenges e. Social equity issues exist in transportation access

[0105] The system identifies the core contradiction: the need for increased mobility vs. the need to reduce traffic and emissions.Step 4: TRIZ Module Activation

[0106] The TRIZ Module analyzes the contradiction and applies relevant inventive principles: a. Principle of Segmentation: Breaking down the transportation system into smaller, more flexible units b. Principle of Dynamization: Making the system more adaptive to changing demandsAttorney Docket No. 68036.2WO01 c. Principle of Self-Service: Empowering users to fulfill their own transportation needs d. Principle of Universality: Creating multi-purpose transportation solutions Step 5: Solution Generation

[0107] The TRIZ Module, in conjunction with the knowledge graph, generates several innovative solution proposals:1. Adaptive Multi-Modal Transit Network: a. A flexible system that dynamically adjusts to demand, combining traditional public transit with on-demand autonomous shuttles. b. Uses Al to predict and meet transportation needs in real-time.2. Elevated Personal Rapid Transit (PRT) System: a. A network of small, electric, autonomous pods that run on elevated guideways above existing roads. b. Provides point-to-point transportation without stops, reducing congestion at street level.3. Integrated Micromobility Hubs: a. Strategically placed hubs throughout the city offering a variety of shared micromobility options (e-bikes, e-scooters, small EVs). b. Seamlessly integrated with public transit and equipped with smart charging infrastructure.4. Gamified Green Commuting Platform: a. A citywide app that incentivizes and rewards sustainable transportation choices. b. Uses social competition and tangible rewards to encourage behavior change. Step 6: Solution Refinement

[0108] The system evaluates these solutions based on feasibility, potential impact, and alignment with the city's specific needs. It then refines the proposals, addressing potential challenges and incorporating successful elements from case studies in the knowledge graph.Step 7: Human- Al Collaboration

[0109] The system presents these refined solutions to the city planner through an interactive interface, providing: a. Visualizations of how each solution would function in the city b. Projected impacts on congestion, emissions, and mobilityAttorney Docket No. 68036.2WO01 c. Estimated implementation costs and timelines d. Potential challenges and mitigation strategies

[0110] The city planner can explore each solution in depth, ask questions, and provide feedback. For instance, they might express concerns about the visual impact of the elevated PRT system. The system would then iterate on this solution, perhaps suggesting alternative designs or routes that minimize visual disruption.Final Proposed Innovative Solution:

[0111] After several rounds of refinement and feedback, the system proposes a comprehensive, multi-faceted solution:"Integrated Smart Mobility Ecosystem"1. Core Adaptive Transit Network: a. Al-driven, demand-responsive public transit system combining buses, trams, and autonomous shuttles. b. Dynamic routing and scheduling to maximize efficiency and coverage.2. Micromobility Integration: a. Network of smart mobility hubs offering e-bikes, e-scooters. and small EVs for last-mile connectivity. b. Seamless integration with the core transit network for easy multi-modal journeys.3. Intelligent Traffic Management: a. Smart traffic lights and sensors to optimize traffic flow and prioritize public transit and emergency vehicles. b. Real-time data sharing with navigation apps to reduce congestion.4. Green Commute Rewards Program: a. Gamified app incentivizing sustainable transportation choices. b. Rewards redeemable for public transit credits, micromobility access, or local business discounts.5. Phased Implementation of Personal Rapid Transit: a. Initial pilot of elevated PRT in high-congestion corridors, with designs sensitive to urban aesthetics. b. Expandable based on success and public acceptance.

[0112] This solution addresses the core contradiction by enhancing mobility options while simultaneously working to reduce congestion and emissions. ItAttorney Docket No. 68036.2WO01 leverages emerging technologies, encourages behavior change, and provides flexibility for future expansion and adaptation.

[0113] The city planner can now use this proposal as a foundation for further discussion with stakeholders, detailed feasibility studies, and potential implementation planning.

[0114] FIG. 6 illustrates a flow chart for a method S 100. in accordance with embodiments of the present disclosure. In one or more embodiments, the method S I 00 may include a method for generating innovative solutions using the system 10 and related disclosures, as described with respect to FIGs. 1-5.

[0115] As shown in FIG. 6, the method S100 includes, at step SI 10, acquiring information from diverse sources using an artificial intelligent (Al) agent, such as the Al agent 100, as described with respect to FIG. 1. The method SI 00 also includes, at step S120, updating a knowledge graph, such as knowledge graph 101 or knowledge graph 30, with the processed information from the Al agent.

[0116] The method S100 further includes, at step S130, identifying crossdomain patterns and relationships within the knowledge graph, as described herein. The method S I 00 further includes, at step SI 40, detecting contradictions within the knowledge graph, as described above with respect to FIGs. 1-5.

[0117] The method SI 00 further includes, at step SI 50, generating solution proposals using a TRIZ Module, such as the TRIZ module 102, that applies TRIZ inventive principles based on the detected contradictions and information from the knowledge graph.

[0118] In one or more embodiments of the method S I 00. the acquiring information may include ingesting data from scientific journals, patent databases, conference proceedings, and technical reports using web scraping and API integration protocols; applying natural language processing techniques including tokenization, named entity recognition, and semantic parsing to extract key concepts; and employing machine learning algorithms including clustering and classification algorithms to identify patterns in the processed information.

[0119] In one or more embodiments of the method S I 00, the natural language processing techniques utilize a transformer architecture based on an enhanced Generative Pre-trained Transformer model fine-tuned for scientific and technical content understanding.Attorney Docket No. 68036.2WO01

[0120] In one or more embodiments of the method S I 00. updating the knowledge graph may include implementing a hypergraph data structure with nodes representing concepts and hyperedges representing multi-way relationships; adaptively creating new nodes for novel concepts identified in the processed information; and dynamically weighting relationships between nodes based on frequency and recency of connections.

[0121] In one or more embodiments of the method S I 00. identifying crossdomain patterns and relationships comprises applying semantic reasoning algorithms that incorporate ontology integration and Resource Description Framework triple stores to identify non-obvious connections between concepts from different technical domains.

[0122] In one or more embodiments of the method S I 00, detecting contradictions may include applying semantic analysis algorithms to identify opposing requirements in textual descriptions; analyzing graph structure to identify nodes with conflicting attributes; and performing time series analysis to detect contradictions in temporal data patterns.

[0123] In one or more embodiments, the method SI 00 may further optionally include, at step SI 60, evaluating generated solution proposals based on novelty, feasibility, and effectiveness criteria; at step 170, refining the solution proposals through iterative feedback analysis; and at step 180, presenting the refined solution proposals through a human-AI collaboration interface that provides interactive visualizations and natural language explanations of the solution derivation process.

[0124] FIG. 7 illustrates a flow chart for a method S200, in accordance with embodiments of the present disclosure. In one or more embodiments, the method S200 may include a method for generating innovative solutions using the system 10 and related disclosures, as described with respect to FIGs. 1-5.

[0125] As shown in FIG. 7, the method S200 includes, at step S210, receiving a problem input from a knowledge graph, such as the knowledge graph 101 or knowledge graph 30, as described herein.

[0126] The method S200 includes, at step S220, performing problem analysis to break down the problem into analyzable components, as described herein. In one or more implementations, the problem is broken down into analyzable components using the TRIZ module 102.Attorney Docket No. 68036.2WO01

[0127] The method S200 includes, at step S230, identifying contradictions within the problem. In one or more embodiments, identifying contradictions includes identifying logical or physical contradictions, with parameters including contradiction types, severity scoring, and context consideration factors.

[0128] The method S200 includes, at step S240, mapping identified contradictions to inventive principles using a contradiction matrix. In one or more embodiments, an output from a contradiction matrix is a prioritized list of applicable inventive principles with relevance scores. In one or more embodiments, an output from inventive principles includes actionable principle descriptions with comprehensive implementation strategies.

[0129] The method S200 includes, at step S250, applying the principles to generate solution concepts; at step S260, creating comprehensive solution packages; at step S270, evaluating the generated solutions; and at step S280, refining the solutions through an iterative process.

[0130] In one or more embodiments of the method S200. the problem analysis includes parameter identification, system analysis, and functional analysis. In one or more embodiments of the method S200, identifying contradictions includes distinguishing between technical contradictions and physical contradictions. In one or more embodiments of the method S200, evaluating the generated solutions includes novelty assessment, feasibility analysis, and effectiveness measurement.

[0131] In accordance with one or more embodiments, a system for generating innovative solutions is provided. The system includes, a literature review Al agent, such as the Al agent 100, configured to analyze information sources; a knowledge graph, such as the knowledge graph 101, configured to organize and connect information; and a TRIZ Module, such as the TRIZ module 102, configured to identify contradictions and generate innovative solutions; wherein the literature review Al agent 100, the knowledge graph, and TRIZ Module are operatively connected to each other.

[0132] In one or more embodiments of the system, the TRIZ Module comprises: a Problem Input component configured to receive structured problem statements from the knowledge graph; a Problem Analysis component configured to break down problems into analyzable components; a Contradiction Identification component configured to identify technical and physical contradictions; a Contradiction Matrix component configured to map identified contradictions toAttorney Docket No. 68036.2WO01 inventive principles; an Inventive Principles component configured to provide implementation guidance for selected principles; a Principle Application component configured to translate principles into solution concepts; a Solution Generation component configured to create comprehensive solution packages; a Solution Evaluation component configured to assess generated solutions; and a Solution Refinement component configured to iteratively improve solutions.

[0133] In one or more embodiments of the system, the Problem Input component is configured to process cross-disciplinary patterns, user-submitted innovation challenges, and automatically identified research gaps.

[0134] In one or more embodiments of the system, the Problem Analysis component is configured to perform parameter identification, system analysis, and functional analysis. In one or more embodiments of the system, the parameter identification maps problem elements to standardized engineering parameters.

[0135] In one or more embodiments of the system, the system analysis identifies super-system, system, and sub-system relationships. In one or more embodiments of the system, the functional analysis examines useful functions, harmful functions, and insufficient functions.

[0136] In one or more embodiments of the system, the Contradiction Identification component is configured to distinguish between technical contradictions and physical contradictions. In one or more embodiments of the system, technical contradictions are identified when improving one parameter leads to degradation of another parameter. In one or more embodiments of the system, physical contradictions are identified when a system requires opposite states of the same parameter. In one or more embodiments of the system, the Contradiction Matrix component is configured to perform matrix lookup procedures, principle selection, and relevance scoring.

[0137] In one or more embodiments of the system, the Inventive Principles component is configured to provide principle definitions, sub-principles, domain applications, and historical examples.

[0138] In one or more embodiments of the system, the Principle Application component is configured to perform contextualization, combination logic, feasibility filtering, and concept development. In one or more embodiments of the system, the Solution Generation component is configured to produce technical specifications, implementation roadmaps, resource requirements, and performanceAttorney Docket No. 68036.2WO01 predictions. In one or more embodiments of the system, the Solution Evaluation component is configured to perform novelty assessment, feasibility analysis, and effectiveness measurement.

[0139] In one or more embodiments of the system, the novelty assessment includes patent landscape analysis, prior art comparison, and innovation scoring.

[0140] In one or more embodiments of the system, the feasibility analysis includes technical readiness evaluation, manufacturing capability assessment, regulatory compliance checks, and market adoption barrier analysis.

[0141] In one or more embodiments of the system, the effectiveness measurement includes performance improvement quantification, cost-benefit analysis, risk assessment, and scalability evaluation.

[0142] In one or more embodiments of the system, the Solution Refinement component is configured to perform gap analysis, parameter adjustment, alternative exploration, re-evaluation, and convergence checking.

[0143] In one or more embodiments, the system further includes a Proposed Innovative Solutions component configured to present final innovation recommendations in comprehensive solution packages. In one or more embodiments, each solution package includes an executive summary, technical specifications, an implementation plan, and a business case.

[0144] In accordance with one or more embodiments, a non-transitory computer-readable medium storing instructions that, when executed by a processor, cause the processor to perform the methods, including methods SI 00 and S200, as disclosed herein.

[0145] In accordance with one or more embodiments, a system for generating innovative solutions is provided. The system includes a literature review Al agent; a knowledge graph; and a TRIZ Module configured to process problem inputs from the knowledge graph, identify contradictions, apply inventive principles, and generate innovative solutions through an iterative refinement process.

[0146] In one or more embodiments of the system, the TRIZ Module is configured to leverage cross-domain knowledge to suggest solutions that bridge different fields of knowledge. In one or more embodiments of the system, the TRIZ Module is configured to incorporate human feedback to refine generated solutions. In one or more embodiments of the system, the TRIZ Module is configured to adapt problem-solving strategies based on updates to the knowledge graph.Attorney Docket No. 68036.2WO01

[0147] In accordance with one or more embodiments, a method for generating innovative solutions using TRIZ principles is provided. The method may include receiving a problem statement with context and constraints; identifying parameters and performing system and functional analysis; detecting technical and physical contradictions; applying relevant inventive principles to resolve the contradictions; generating solution concepts based on the applied principles; evaluating the solutions based on novelty, feasibility, and effectiveness; and presenting refined innovative solutions with implementation details. In one or more embodiments of the method, the problem statement is derived from cross-disciplinary patterns detected in a knowledge graph.

[0148] In accordance with one or more embodiments, a method is provided. The method includes receiving, via a graphical user interface of an artificial intelligence (Al) agent, a query associated with a topic in a knowledge domain; decoding, via the Al agent, the query to identify relevant technical information associated with the topic; acquiring, via the Al agent, data associated with the relevant technical information from a database; analyzing, via the Al agent, the acquired data to generate one or more key features associated with the query; extracting the one or more identified key features to a knowledge graph; and updating the knowledge graph based on the one or more extracted key features to organize information in representative patterns and relationships.

[0149] In one or more embodiments of the method, the database may include scientific journals, patent databases, conference proceedings, technical reports, social media feeds, news articles, academic theses, and a combination thereof. In one or more embodiments of the method, the one or more key features relate to key concept identification, relationship mapping, hypothesis formulation, or a combination thereof. In one or more embodiments of the method, the knowledge graph may include a cross-disciplinary7knowledge graph that represents interconnections between various scientific disciplines, concepts, and methodologies.

[0150] In accordance with one or more embodiments, a method is provided. The method includes receiving, via a knowledge graph of an artificial intelligence (Al) agent, a structured problem statement related to a knowledge domain; analyzing the structured problem statement via the Al agent to generate a comprehensive system model comprising identified parameters, functions, andAttorney Docket No. 68036.2WO01 system hierarchy, wherein the analyzing the structured problem statement further comprises performing parameter identification on the identified parameters, and performing functional analysis of the comprehensive system model; identifying, via the Al agent, a contradiction generated by the functional analysis; resolving, via a contradiction matrix of the Al agent, the contradiction as a technical contradiction or a physical contradiction; proposing an innovative solution based on resolved contradiction.

[0151] In one or more embodiments of the method, the request may include a request statement, context, and constraints. In one or more embodiments of the method, the knowledge graph may include a cross-disciplinary knowledge graph that represents interconnections between various scientific disciplines, concepts, and methodologies.

[0152] In one or more embodiments of the method, the technical contradiction is identified when improving one parameter leads to degradation of another parameter and a physical contradiction is identified when a system requires opposite states of the same parameter.

[0153] FIG. 8 is a simplified diagram illustrating a neural network structure that may be implemented by one or more components in the artificial intelligence (Al) agent 100, the knowledge graph 101, and the TRIZ module 102, according to some embodiments. As used herein, the neural network structure may comprise any hardware or software-based framework that includes any artificial intelligence network or system, neural network or system and / or any training or learning models implemented thereon or therewith. The neural network comprises a computing system that is built on a collection of connected units or nodes, referred to as neurons (e.g., 844, 845, 846). Neurons are often connected by edges. Each neuron may associate an adjustable weight (e.g., 851, 852) with the edge. The neurons may be aggregated into layers such that different layers may perform different transformations on the respective input and output transformed input data to the next layer.

[0154] For example, the neural network architecture may comprise an input layer 841, one or more hidden layers 842 and an output layer 843. Each layer may comprise a plurality of neurons, and neurons between layers are interconnected according to a specific topology of the neural network topology. The input layer 841 receives the input data, such as data associated with nodes or edges inAttorney Docket No. 68036.2WO01 knowledge graph 101. The number of nodes (neurons) in the input layer 841 may be determined by the dimensionality of the input data (e.g., the length of a vector of give an example of the input). Each node in the input layer represents a feature or attribute of the input.

[0155] The hidden layers 842 are intermediate layers between the input and output layers of a neural network. It is noted that two hidden layers 842 are shown in FIG. 8 for illustrative purpose only, and any number of hidden layers may be utilized in a neural network structure. Hidden layers 842 may extract and transform the input data through a series of weighted computations and activation functions.

[0156] For example, the neural network structure that is embedder 220 receives an input of nodes and / or edges and transforms the input into an output that is a vector embedding. In some instances, the input layer 841 or a combination of input layer 841 and a subset of hidden layers 842 may also function as an encoder. To perform the transformation, each neuron receives input signals, performs a weighted sum of the inputs according to weights assigned to each connection (e.g., 851. 852), and then applies an activation function (e.g., 861. 862. etc.) associated with the respective neuron to the result. The output of the neuron is passed to the next layer of neurons or serves as the final output of the network. The activation function may be the same or different across different layers. Example activation functions include but not limited to Sigmoid, hyperbolic tangent, Rectified Linear Unit (ReLU), Leaky ReLU, Softmax, and / or the like. In this way, after a number of hidden layers, input data received at the input layer 841 is transformed into rather different values indicative data characteristics corresponding to a task that the neural network structure has been designed to perform.

[0157] The output layer 843 is the final layer of the neural network structure. It produces the network's output or prediction based on the computations performed in the preceding layers (e.g., 841, 842). The number of nodes in the output layer depends on the nature of the task being addressed. For example, in a binary classification problem, the output layer may consist of a single node representing the probability of belonging to one class. In a multi-class classification problem, the output layer may have multiple nodes, each representing the probability of belonging to a specific class.

[0158] Therefore, the machine learning model(s) in the artificial intelligence (Al) agent 100, the knowledge graph 101, and the TRIZ module 102 may compriseAttorney Docket No. 68036.2WO01 the transformative neural network structure of layers of neurons, and weights and activation functions describing the non-linear transformation at each neuron. Such a neural network structure is often implemented on one or more hardware processors (e.g., processing component 904), such as a graphics processing unit (GPU). An example neural network may include logistic regression, random tress, multi-layer perceptron, and / or the like.

[0159] In one embodiment, the machine learning model(s) in the artificial intelligence (Al) agent 100, the knowledge graph 101, and the TRIZ module 102 may be implemented by hardware, software and / or a combination thereof. For example, the the artificial intelligence (Al) agent 100, the knowledge graph 101, and the TRIZ module 102 may comprise a specific neural network structure implemented and run on various hardware platforms 860, such as but not limited to CPUs (central processing units), GPUs (graphics processing units), FPGAs (field- programmable gate arrays), Application-Specific Integrated Circuits (ASICs), dedicated Al accelerators like TPUs (tensor processing units), and specialized hardware accelerators designed specifically for the neural network computations described herein, and / or the like. Example specific hardware for neural network structures may include, but not limited to Google Edge TPU, Deep Learning Accelerator (DLA), NVIDIA Al-focused GPUs, and / or the like. The hardware 860 used to implement the neural netw ork structure is specifically configured based on factors such as the complexity of the neural network, the scale of the tasks (e.g., training time, input data scale, size of training dataset, etc.), and the desired performance.

[0160] In one embodiment, a neural network-based artificial intelligence (Al) agent 100, knowledge graph 101, and TRIZ module 102 may be trained by iteratively updating the underlying parameters (e.g., weights 851, 852, etc., bias parameters and / or coefficients in the activation functions 861, 862 associated with neurons) of the neural network based on a loss function, such as a mean squared estimation error (MSEE). For example, during forward propagation, the training data such as vector representations of encoded nodes are fed into the neural network. The data flows through the network's layers 841, 842, with each layer performing computations based on its weights, biases, and activation functions until the output layer 843 produces the network's output. In some embodiments, outputAttorney Docket No. 68036.2WO01 layer 843 produces an intermediate output on which the network’s final output is based.

[0161] The output generated by the output layer 843 is compared to the expected output (e.g., a “ground-truth” such as the ground truth node classifications) from the training data, to compute a loss function that measures the discrepancy between the predicted output and the expected output. For example, the loss function may be MSEE. Given the loss, the negative gradient of the loss function is computed with respect to each weight of each layer individually. Such negative gradient is computed one layer at a time, iteratively backward from the output layer 843 to the input layer 841 of the neural network. These gradients quantify the sensitivity of the network's output to changes in the parameters. The chain rule of calculus is applied to efficiently calculate these gradients by propagating the gradients backward from the output layer 843 to the input layer 841.

[0162] Parameters of the neural network are updated backwardly from the last layer to the input layer (backpropagating) based on the computed negative gradient using an optimization algorithm to minimize the loss. The backpropagation from the output layer 843 to the input layer 841 may be conducted for a number of training samples in a number of iterative training epochs. In this way, parameters of the neural network may be gradually updated in a direction to result in a lesser or minimized loss, indicating the neural network has been trained to generate a predicted output value closer to the target output value with improved prediction accuracy. Training may continue until a stopping criterion is met, such as reaching a maximum number of epochs or achieving satisfactory performance on the validation data. At this point, the trained network can be used to make predictions on new, unseen data, such as generating a new edge from a different graph projection.

[0163] Neural network parameters may be trained over multiple stages. For example, initial training (e.g., pre-training) may be performed on one set of training data, and then an additional training stage (e.g., fine-tuning) may be performed using a different set of training data. In some embodiments, all or a portion of parameters of one or more neural-network model being used together may be frozen, such that the “frozen” parameters are not updated during that training phase.Attorney Docket No. 68036.2WO01This may allow, for example, a smaller subset of the parameters to be trained without the computing cost of updating all of the parameters.

[0164] Therefore, the training process transforms the neural network into an “updated” trained neural network with updated parameters such as weights, activation functions, and biases. The trained neural network thus improves neural network technology in protein prediction.

[0165] FIG. 9 is a block diagram of a computer system 900 suitable for implementing various methods and devices described in FIGs. 1-8.

[0166] In accordance with various aspects of the present disclosure, the computer system 900. such as a network server or a mobile communications device, may include a bus component 902 or other communication mechanisms for communicating information, which interconnects subsystems and components, such as a computer processing component 904 (e.g., processor, micro-controller, digital signal processor (DSP), etc.), system memory component 906 (e.g., RAM), static storage component 908 (e.g.. ROM), disk drive component 910 (e.g., magnetic or optical), network interface component 912 (e.g.. modem or Ethernet card), display component 914 (e.g., cathode ray tube (CRT) or liquid crystal display (LCD)), input component 916 (e.g., keyboard), cursor control component 918 (e.g., mouse or trackball), and image capture component 920 (e g., analog or digital camera). In one implementation, disk drive component 910 may comprise a database having one or more disk drive components.

[0167] In accordance with aspects of the present disclosure, computer system 900 performs specific operations by the processing component 904 executing one or more sequences of one or more instructions contained in system memory component 906. Such instructions may be read into system memory component 906 from another computer readable medium, such as static storage component 908 or disk drive component 910. In other aspects, hard-wired circuitry may be used in place of (or in combination with) software instructions to implement the present disclosure. In some aspects, the various components of the artificial intelligence (Al) agent 100, the knowledge graph 101, and the TRIZ module 102 may be in the form of software instructions that can be executed by the processing component 904 to automatically perform context-appropriate tasks on behalf of a user.

[0168] Logic may be encoded in a computer readable medium, which may refer to any medium that participates in providing instructions to the processingAttorney Docket No. 68036.2WO01 component 904 for execution. Such a medium may take many forms, including but not limited to, non-volatile media and volatile media. In one aspect, the computer readable medium is non-transitory. In various implementations, non-volatile media includes optical or magnetic disks, such as disk drive component 910, and volatile media includes dynamic memory, such as system memory component 906. In one aspect, data and information related to execution instructions may be transmitted to computer system 900 via a transmission media, such as in the form of acoustic or light waves, including those generated during radio wave and infrared data communications. In various implementations, transmission media may include coaxial cables, copper wire, and fiber optics, including wires that comprise bus 902.

[0169] Some common forms of computer readable media include, for example, floppy disk, flexible disk, hard disk, magnetic tape, any other magnetic medium, CD-ROM, any other optical medium, punch cards, paper tape, any other physical medium with patterns of holes, RAM. PROM, EPROM, FLASH-EPROM, any other memory’ chip or cartridge, carrier wave, or any other medium from which a computer is adapted to read. These computer readable media may also be used to store the programming code for the artificial intelligence (Al) agent 100, the knowledge graph 101, and the TRIZ module 102 discussed above.

[0170] In various aspects of the present disclosure, execution of instruction sequences to practice the present disclosure may be performed by computer system 900. In various other aspects of the present disclosure, a plurality of computer systems 900 coupled by communication link 930 (e.g., a communications network, such as a LAN, WLAN, PTSN, and / or various other wired or wireless networks, including telecommunications, mobile, and cellular phone networks) may perform instruction sequences to practice the present disclosure in coordination with one another.

[0171] Computer system 900 may transmit and receive messages, data, information and instructions, including one or more programs (i.e., application code) through communication link 930 and network interface component 912. Received program code may be executed by processing component 904 as received and / or stored in disk drive component 910 or some other non-volatile storage component for execution. The communication link 930 and / or the network interface component 912 may be used to conduct electronic communications between the artificial intelligence (Al) agent 100, the knowledge graph 101, and the TRIZAttorney Docket No. 68036.2WO01 module 102, as described with respect to FIG. 1, and external devices, for example with the device 110 depending on where the artificial intelligence (Al) agent 100, the knowledge graph 101, and the TRIZ module 102 may be implemented.

[0172] Where applicable, various embodiments provided by the disclosure may be implemented using hardware, software, or combinations of hardware and software. Also, where applicable, the various hardware components and / or software components set forth herein may be combined into composite components comprising software, hardware, and / or both without departing from the scope of the disclosure. Where applicable, the various hardware components and / or software components set forth herein may be separated into sub-components comprising software, hardware, or both without departing from the scope of the disclosure. In addition, where applicable, it is contemplated that software components may be implemented as hardware components and vice-versa.

[0173] Software, in accordance with the disclosure, such as program code and / or data, may be stored on one or more computer readable mediums. It is also contemplated that software identified herein may be implemented using one or more general purpose or specific purpose computers and / or computer systems, networked and / or otherwise. Where applicable, the ordering of various steps described herein may be changed, combined into composite steps, and / or separated into sub-steps to provide features described herein.

[0174] The foregoing outlines features of several embodiments so that those skilled in the art may better understand the aspects of the present disclosure. Those skilled in the art should appreciate that they may readily use the present disclosure as a basis for designing or modifying other processes and structures for carrying out the same purposes and / or achieving the same advantages of the embodiments introduced herein. Those skilled in the art should also realize that such equivalent constructions do not depart from the spirit and scope of the present disclosure, and that they may make various changes, substitutions, and alterations herein without departing from the spirit and scope of the present disclosure.

Claims

Attorney Docket No. 68036.2WO01WHAT IS CLAIMED IS;1. An integrated system for generating innovative solutions, comprising: an artificial intelligent (Al) agent configured to continuously analyze diverse information sources and extract structured data; a knowledge graph operatively connected to the Al agent and configured to organize and connect information received from the Al agent, wherein the knowledge graph identifies patterns and relationships across different domains; and a TRIZ Module operatively connected to the knowledge graph and configured to identify contradictions within the knowledge graph and generate innovative solutions by applying TRIZ principles based on information from the knowledge graph.

2. The integrated system of claim 1, wherein the Al agent comprises: a multi-source data ingestion module configured to acquire information from scientific journals, patent databases, and technical publications; a natural language processing system configured to extract key concepts and relationships from textual content; and a machine learning algorithm suite configured to identify patterns and trends in the analyzed information.

3. The integrated system of claim 2, wherein the natural language processing system employs a transformer architecture based on an enhanced Generative Pre-trained Transformer model fine-tuned for scientific and technical content understanding.

4. The integrated system of claim 1, wherein the knowledge graph comprises: a hypergraph data structure including nodes representing concepts, edges representing relationships between concepts, and hyperedges representing complex multi-way relationships; a self-organizing structure mechanism configured to adaptively create new nodes and dynamically weight relationships based on incoming information; and a semantic reasoning engine configured to identify cross-domain connections and generate hypotheses.Attorney Docket No. 68036.2WO015. The integrated system of claim 4, wherein the semantic reasoning engine incorporates ontology integration, a Resource Description Framework triple store, and a SPARQL query engine for complex semantic queries.

6. The integrated system of claim 1, wherein the TRIZ Module comprises: an automated contradiction identification system configured to detect technical and physical contradictions using semantic analysis and graph-based detection algorithms; a TRIZ inventive principle application mechanism configured to apply inventive principles through a principle mapping neural network; and a cross-domain solution suggestion engine configured to identify analogous solutions from different technical domains.

7. The integrated system of claim 6, wherein the cross-domain solution suggestion engine includes an analogical reasoning engine configured to identify structurally similar problems across different domains and a solution transfer algorithm configured to adapt solutions from analogous problems to fit current problem contexts.

8. A method for generating innovative solutions, comprising: acquiring information from diverse sources using an artificial intelligent (Al) agent; updating a knowledge graph with the processed information from the Al agent; identifying cross-domain patterns and relationships within the knowledge graph; detecting contradictions within the knowledge graph; and generating solution proposals using a TRIZ Module that applies TRIZ inventive principles based on the detected contradictions and information from the knowledge graph.

9. The method of claim 8, wherein acquiring information comprises: ingesting data from scientific j oumals, patent databases, conference proceedings, and technical reports using web scraping and API integration protocols; applying natural language processing techniques including tokenization, named entity recognition, and semantic parsing to extract key concepts; and employing machine learning algorithms including clustering and classification algorithms to identify patterns in the processed information.Attorney Docket No. 68036.2WO0110. The method of claim 9, wherein the natural language processing techniques utilize a transformer architecture based on an enhanced Generative Pre-trained Transformer model fine-tuned for scientific and technical content understanding.

11. The method of claim 8, wherein updating the knowledge graph comprises: implementing a hypergraph data structure with nodes representing concepts and hyperedges representing multi-way relationships; adaptively creating new nodes for novel concepts identified in the processed information; and dynamically weighting relationships between nodes based on frequency and recency of connections.

12. The method of claim 11, wherein identifying cross-domain patterns and relationships comprises applying semantic reasoning algorithms that incorporate ontology integration and Resource Description Framework triple stores to identify non-obvious connections between concepts from different technical domains.

13. The method of claim 8, wherein detecting contradictions comprises: applying semantic analysis algorithms to identify opposing requirements in textual descriptions; analyzing graph structure to identify nodes with conflicting attributes; and performing time series analysis to detect contradictions in temporal data patterns.

14. The method of claim 8, further comprising: evaluating generated solution proposals based on novelty, feasibility, and effectiveness criteria; refining the solution proposals through iterative feedback analysis; and presenting the refined solution proposals through a human- Al collaboration interface that provides interactive visualizations and natural language explanations of the solution derivation process.

15. A computer-readable medium storing instructions that, when executed by a processor, cause the processor to perform operations comprising: operating an Al agent to continuously scan and analyze information sources;Attorney Docket No. 68036.2WO01 maintaining a knowledge graph that receives and organizes information from the Al agent into interconnected nodes and relationships; and executing a TRIZ Module that accesses the knowledge graph to identify technical contradictions and generate innovative solutions by applying systematic problem-solving principles.

16. The computer-readable medium of claim 15, wherein operating the Literature Review Al Agent comprises: executing a multi-source data ingestion module to acquire information from scientific journals, patent databases, conference proceedings, and technical reports; applying natural language processing algorithms including tokenization, named entity recognition, and semantic parsing to extract key concepts from textual content; and employing machine learning algorithms including clustering and classification algorithms to identify' patterns and trends in the analyzed information.

17. The computer-readable medium of claim 16, wherein the natural language processing algorithms utilize a transformer architecture based on an enhanced Generative Pre-trained Transformer model fine-tuned for scientific and technical content understanding.

18. The computer-readable medium of claim 15, wherein maintaining the knowledge graph comprises: implementing a hypergraph data structure with nodes representing concepts, edges representing binary relationships, and hyperedges representing complex multi-way relationships; adaptively creating new nodes for novel concepts identified in processed information; and dynamically weighting relationships between nodes based on frequency and recency of connections.

19. The computer-readable medium of claim 18, wherein the hypergraph data structure incorporates a semantic reasoning engine that includes ontology integration, a Resource Description Framework triple store, and a SPARQL query' engine for performing complex semantic queries across the graph structure.Attorney Docket No. 68036.2WO0120. The computer-readable medium of claim 15, wherein executing the TRIZ Module comprises: applying semantic analysis algorithms and graph-based detection algorithms to automatically identify technical and physical contradictions within the knowledge graph; utilizing a principle mapping neural network to select and apply appropriate TRIZ inventive principles based on identified contradiction characteristics; and generating solution proposals through an analogical reasoning engine that identifies structurally similar problems across different technical domains within the knowledge graph.

Images