129 results about "Heuristic" patented technology

The present invention discloses a modular AI-based system for knowledge instantiation and evidence synthesis, unifying five modules: Knowledge Instantiation, ingesting anchor knowledge via retrieval- augmented generation and neuromorphic prompting; Case-Based Agent Generation, orchestrating domain constraints, performance criteria, and sub-task decomposition; Secondary Epistemogenesis, spawning sub-agents with inherited heuristics; Workflow Orchestration and Evidence Synthesis, merging outputs, tracking metadata, and storing final artifacts; and Reinforcement Learning from Human Refinement, capturing feedback and edits. The method for output generation initiates by defining knowledge requirements, indexing public and proprietary knowledge, validating output structure, drafting content, and refining outputs via sub-agent spawning and human oversight. The system ensures domain alignment, concurrency control, and persistent versioning. By combining specialized knowledge sources, multi-agent concurrency, and iterative human feedback loops, it dynamically adapts to evolving knowledge requirements. This invention emphasizes reliability, traceability, and AI-driven outputs in regulated industries, delivering trustworthy outcomes anchored in expert knowledge.

The invention relates to the field of automatic algorithm design, in particular to an automatic algorithm design method for solving a vehicle path planning problem by utilizing a large language model, which comprises the following steps of: S1, constructing an AutoDH framework, and defining a decision form of an intelligent agent; s2, defining heuristics in an LLM pool, an improvement pool and a disturbance pool; s3, acquiring state information of the CVRPs, and intelligently selecting a heuristic mode through two-stage MDP; s4, designing a reward mechanism fusing solution quality improvement, heuristic time cost and LLMs API calling cost; and S5, training the intelligent agent to optimize the solution of the CVRPs. The method can intelligently select the most cost-effective heuristic algorithm according to the state information of the current CVRPs, and improves the flexibility and adaptability of algorithm design.

A system, method, and device for adaptively limiting web requests is disclosed. The method includes (i) receiving a request at a proxy for an authentication service, (ii) determining whether the request is trusted based at least in part on a classification obtained from a classifier, and (iii) handling the request according to a determination of whether the request is trusted.

A neural network on-chip mapping method and apparatus based on a tabu search algorithm are provided. The method includes: constructing a tabu search table and using a heuristic-based iterative search process to select local computing cores of a network-on-chip as candidates, establishing an integer programming model and solving an optimal solution, continuously reducing an objective cost function of a mapping solution by loop iteration, and finally obtaining an approximately optimal deployment scheme.

Systems and methods for training and using large language models to respond to queries and cascading to new models when their performance falls below a threshold during the response generation cycle are described. The methods involve receiving and analyzing a query using heuristics to determine the query's categories and a level of granularity at which its response is to be evaluated. Selecting a language model based on the query analysis and during the generation of the response, evaluating the model's performance for its ability to predict the next segment in a response. The methods score the evaluation and use the score to determine whether the language model's performance exceeds a confidence threshold. If it does not, then during the response is being generated, e.g., at a point before the response generation is completed, cascading from the model to a different model that can perform at a higher confidence level.

Disclosed apparatuses and methods for trajectory and collision prediction comprising a processor operable to generate, using a machine-learning (ML) model, a ML predicted trajectory of a vehicle and a ML model confidence of the ML predicted trajectory, generate, using a heuristics model, a heuristic predicted trajectory of the vehicle, determine whether the ML model confidence satisfies a metric, in response to determining that the ML model confidence satisfies the metric, apply the ML predicted trajectory as a predicted trajectory, and in response to determining that the ML learning model confidence fails the metric, apply the heuristic predicted trajectory as the predicted trajectory.

Computerized methodologies are disclosed that are directed to detecting anomalies within a time-series data set. A first aspect of the anomaly detection process includes analyzing the regularity of the data points of the time-series data set and determining whether a data aggregation process is to be performed based on the regularity of the data points, which results in a time-series data set having data points occurring at regular intervals. A seasonality pattern may be determined for the time-series data set, where a silhouette score is computed to measure the quality of the fit of the seasonality pattern to the time-series data. The silhouette score may be compared to a threshold and based on the comparison, the seasonality pattern or a set of heuristics may be utilized in an anomaly detection process. When the seasonality pattern is utilized, the seasonality pattern may be utilized to generate thresholds indicating anomalous behavior.

The invention discloses a video question answering method for specific domain fine-grained heuristic prompt. The method comprises the following steps: 1, designing a prompt template of a specific domain through data engineering; performing space-time cutting processing on the original video data; the space-time cutting treatment is carried out; 2, taking the prompt template and the clipped video data as input of an entity prompter and an action prompter to obtain a heuristic label; and 3, through a heuristic enhanced answer reasoning scheme, collaborative interaction between the visual language basic model and the entity prompter and between the visual language basic model and the action prompter is enhanced, that is, a label output by the visual language basic model is compared with a heuristic label, and the ability of the visual language basic model to accurately deduce answers is improved. According to the method, a large-scale model is converted into a dynamic knowledge engine, complex content can be efficiently and accurately analyzed, and attention to basic cross-modal elements is enhanced by integrating a fine-grained inspiration method in a specific field into design of a prompt template.

The embodiment of the invention discloses a commodity searching method and electronic equipment, and the method comprises the steps that a commodity searching interaction interface is provided, the interface comprises a first area and a second area, the first area is used for providing an artificial intelligence AI interaction component so as to receive a search demand expressed by a user in a mode of inputting a natural language statement, and the second area is used for providing a second area; through multiple rounds of interaction with an AI large model, receiving a statement which is input by the user under the heuristic of the reply content of the AI large model and is used for clearly expressing the search demand of the user; the second area is used for providing a plurality of attribute options for defining search requirements and a plurality of attribute value options corresponding to the attribute options according to interaction conditions in the first area so as to receive search requirement information expressed by the user in a mode of selecting attributes and attribute values; and in each round of interaction, carrying out information fusion on the input statements collected through the first region and the attributes and attribute value selection results collected through the second region so as to provide corresponding commodity search results according to information fusion results. Through the embodiment of the invention, the user can be helped to express the search demand, and the search efficiency is improved.

A training optimization system implements algorithmic solutions to solve the conjoined problem of accelerator architecture search and model partitioning for distributed training. The system makes the multi-dimensional optimization space of architecture search and device placement tractable by reducing the number of accelerator architectures explored through area-based heuristics and employing a novel integer linear program (ILP), the size of which is dependent only on the number of operators. The ILP scheduling optimization also explores the partitioning of operators across cores, known as intra-operator parallelism. Despite the vast space, the ILP described herein requires significantly less time to perform the optimizations across all explored accelerator configurations. Based on the optimal backward and forward pass latencies, the system leverages a novel dynamic programming (DP) approach to determine the device placement and model partitioning scheme.

A data platform monitors a compute environment by performing multi-stage heuristic analysis of event data representing a plurality of events occurring within the environment. The platform utilizes multiple event analyzers, each configured according to a distinct analysis heuristic, to evaluate different subsets of the event data and generate corresponding output signals. A higher-level event analyzer applies a further heuristic to the multiple output signals to generate a composite alert signal, indicating whether the combination of analyzed events collectively represents a security intrusion or other anomalous condition of sufficient severity to warrant alerting. Based on the composite alert signal, the platform performs an alert-based operation, such as generating a user-facing alert, initiating an automated mitigation, or updating a contextual model of system behavior. By combining the analytical outputs of heterogeneous heuristics, the disclosed architecture enhances the accuracy and contextual relevance of automated intrusion detection within complex computing environments.

This disclosure describes techniques and architecture provide automated contribution analysis for “why question” style NLQ answering, e.g., “why is revenue down in North America Q1 2022.” In particular, the techniques described herein combine multiple signals together including, for example, frequency of use of combinations of dimensions in previous NLQs (warm-start), statistical information about columns (e.g., entropy), correlation / co-occurrence between pairs of dimension columns, and correlation between dimensions and dates. This information is used with a set of heuristics and rules to pick the best set of dimensions as contributing factors for a particular metric over a particular time period and present an automatic contribution analysis to the users to give them insights into their data.

Systems and Methods for Automatically Detecting Root Cause Failures Described herein are techniques for automatically detecting root cause failures in a computing environment. A data center may experience a critical system failure that renders the data center inoperable. An administrator may utilize a root cause detector to analyze alerts generated from the data center to automatically detect the root cause of the critical system failure. Once detected, the administrator may investigate the device with the root cause failure to repair the data center. In some examples, the root cause detector may compare alerts received from the data center with root cause failures in a failure repository that it is familiar with to determine whether the present sequence of alerts is similar to alert patterns it has seen before. The root cause detector may be implemented with a classifier model, a large language model, a rule-based heuristics identifying spurious alert patterns, or combinations of these techniques.

There is provided a system and methods to determine an improved solution to a Mixed Integer Problem (MIP) using a quantum-assisted MIP solver. The methods are performed by a digital processor in communication with a quantum processor. Methods include: selecting at least one feasible solution determined by an MIP solver, determining a first sub-problem of the MIP based on the at least one feasible solution; casting the first sub-problem as Binary Quadratic Models (BQMs); solving the BQMs using the quantum processor to generate sample solutions; determining a second sub-problem based on at least the sample solutions, and obtaining a current solution to the MIP by evaluating the second sub-problem; and updating an incumbent solution if the current solution improves over the current incumbent solution. The quantum-assisted MIP solver uses hybrid crossover and mutation heuristics to improve the convergence time and accuracy of solutions obtained using Branch-and-Cut solvers.

The invention relates to the technical field of SAR (Synthetic Aperture Radar) remote sensing imaging, in particular to a scattering heuristic SAR tomography method for structured three-dimensional reconstruction, which comprises the following steps of: S1, acquiring and preprocessing multi-baseline SAR data; s2, inverting scattering parameters based on an attribute scattering center model and orthogonal matching pursuit; s3, segmenting a building area according to coherence and morphological indication; s4, combining direction perception expansion with region updating, determining a continuous scattering region, and giving reference height prior; and S5, constructing a regional adaptive dictionary, sparsely solving the elevation and geocoding to generate a structured point cloud. According to the method, physical prior and regional statistics are cooperated, geometric continuity and anti-noise robustness are improved, pseudo peaks and calculation redundancy are reduced, and upgrading from sparse point cloud to a structured model is realized.

The invention discloses a question answering method and a related device, and relates to the technical field of large language models.According to the question answering method, a large language model which is obtained through pre-training and has the capacity of discovering the association between a question to be answered and similar questions and utilizing the similar questions for answering serves as an answering model, and after the question to be answered is obtained, the question to be answered is answered through the association between the question to be answered and the similar questions; the method comprises the steps of determining similar questions corresponding to questions and analysis of each similar question, calling a trained answering model to answer to-be-answered questions based on the similar questions and the analysis of the similar questions for each similar question, and obtaining answers output by the answering model; and finally, determining a final answer of the question based on each answer output by the answer model. According to the scheme, the model can give answers by referring to the answering process of the similar questions or obtain'heuristic 'from the answering thought of the similar questions to solve the difficult problem that the questions cannot be answered during direct answering, so that the diversity of the answers generated by the model is improved, and the question answering accuracy is improved.

Described herein are techniques for automatically detecting root cause failures in a computing environment. A data center may experience a critical system failure that renders the data center inoperable. An administrator may utilize a root cause detector to analyze alerts generated from the data center to automatically detect the root cause of the critical system failure. Once detected, the administrator may investigate the device with the root cause failure to repair the data center. In some examples, the root cause detector may compare alerts received from the data center with root cause failures in a failure repository that it is familiar with to determine whether the present sequence of alerts is similar to alert patterns it has seen before. The root cause detector may be implemented with a classifier model, a large language model, a rule-based heuristics identifying spurious alert patterns, or combinations of these techniques.

The invention relates to the technical field of new energy power prediction, and discloses a new energy power prediction method and system. According to the method, through multi-source data preprocessing and correlation screening, key influence factors are focused to construct an initial prediction model; a dynamic mixed meta-heuristic optimization strategy is adopted, global efficient exploration is carried out on the upper layer model, local fine optimization is carried out on the lower layer model, collaborative optimization is carried out on hyper-parameters of the initial prediction model in stages, and finally a target prediction model is formed and used for real-time power prediction. According to the method, through the synergistic effect of the upper-layer model and the lower-layer model of dynamic mixed meta-heuristic optimization, the advantages of different algorithms are combined, and it is ensured that the hyper-parameter optimization process is more likely to converge to a globally optimal solution. The technical problems that an existing single optimization algorithm often has limitation and is prone to falling into local optimum or low in convergence speed are solved.

In some cases, images and depth maps can define bins with objects in random configurations. It is recognized herein that current approaches to training deep neural networks to perform grasp computations lack capabilities and efficiencies, such that the resulting grasp computations and grasps can be imprecise or cumbersome, among other shortcomings. Synthetic depth images can be labeled with grasp annotations that are generated based on heuristic-based analyses, so as to define annotated synthetic datasets. The annotated synthetic datasets can be used to train neural networks to determine the best grasp locations for different objects arranged in a variety of positions with respect to each other.

A process, system and medium for detecting anomalous authentication requests to a protected resource during an authentication journey, in order to regulate step-up authentication are described. The process includes obtaining request features from the authentication request that triggered the authentication journey. The process includes processing, by an ensemble of Machine Learning (ML) models and a set of rule-based heuristics, a set of features based on the request features, the set of features associated with a userID. The process includes deriving risk sub-scores for each ML model and each heuristic. The process includes deriving a risk score based on the risk sub-scores. The process includes determining that the risk score exceeds an explanation-triggering threshold. The process includes providing, to a node in the authentication journey, the risk score with an explanation of the risk score. The system and medium are configured to execute the process, and configured to provide the explanation.

The present invention discloses a source code vulnerability detection method based on contrastive learning. The method generates required sample pairs by customizing semantic-preserving heuristics, represents vulnerability codes from the perspective of code tokens to statements to graphs, and uses code attribute graphs to better represent the deep semantic information of vulnerabilities, fully explore the relationship between vulnerability codes and contexts, adopts contrastive learning technology, performs training on small samples, and realizes the high-precision requirements of small sample data. The present invention correspondingly provides a source code vulnerability detection system based on contrastive learning.

The described techniques provide for an automated process of hardware design using main applications (e.g., use cases) and a modeling heuristic of a compiler. Hardware may be optimized and designed based on the knowledge of the compiler for the hardware and firmware information. For instance, a user may define constraints (e.g., area constraints, power constraints, performance constraints, accuracy degradation constraints, etc.) and the changes of compilation heuristics and hardware parameters may be optimized to efficiently achieve the user defined constraints. Accordingly, hardware configuration parameters may be optimized based on the neural network's compilation process (e.g., actual compiler constraints) and optimization of power, performance, and area (PPA) constraints (e.g., user defined constraints). Specific neural processor (SNP) hardware may thus be designed based on the optimized hardware configuration parameters (e.g., via modeling heuristics of its compiler and main applications informed via user defined constraints or PPA constraints).

A training optimization system implements algorithmic solutions to solve the conjoined problem of accelerator architecture search and model partitioning for distributed training. The system makes the multi-dimensional optimization space of architecture search and device placement tractable by reducing the number of accelerator architectures explored through area-based heuristics and employing a novel integer linear program (ILP), the size of which is dependent only on the number of operators. The ILP scheduling optimization also explores the partitioning of operators across cores, known as intra-operator parallelism. Despite the vast space, the ILP described herein requires significantly less time to perform the optimizations across all explored accelerator configurations. Based on the optimal backward and forward pass latencies, the system leverages a novel dynamic programming (DP) approach to determine the device placement and model partitioning scheme.

Aspects of this technical solution can generate, according to one or more first environment metrics, a three-dimensional (3D) model including a first surface corresponding to one or more physical ways through a physical environment, the one or more first environment metrics indicative of boundaries of the one or more physical ways, generate, according to one or more second environment metrics, one or more geometric two-dimensional (2D) objects on the first surface, the second environment metrics indicative of the one or more physical ways, identify, according to one or more viewpoint metrics indicative of cameras of a physical object configured to move along the one or more physical ways, one or more viewpoints oriented to capture corresponding portions of the 3D model, and render, from the one or more corresponding portions of the 3D model, one or more 2D images each corresponding to respective ones of the viewpoints.

Various embodiments include methods and neural network computing devices implementing the methods for performing quantization in neural networks. Various embodiments may include equalizing ranges of weight tensors or output channel weights within a first layer of the neural network by scaling each of the output channel weights of the first layer by a corresponding scaling factor, and scaling each of a second adjacent layer's corresponding input channel weights by applying an inverse of the corresponding scaling factor to the input channel weights. The corresponding scaling factor may be determined using a black-box optimizer on a quantization error metric or based on heuristics, equalization of dynamic ranges, equalization of range extrema (minima or maxima), differential learning using straight through estimator (STE) methods and a local or global loss, or using an error metric for the quantization error and a black-box optimizer that minimizes the error metric with respect to the scaling.

Systems and techniques for generation and curation of a professional headshot from a set of image data. The systems and techniques images from the set of image data based on characteristics of the representation of the individual within the image. The systems and techniques further include determining a bounding box to define a headshot, the bounding box determined based on guidelines established by heuristics and / or machine learning algorithms trained using data labeled based on heuristics.

The invention discloses an NAS hyperspectral image crop classification method combined with noise inspiration and attention mechanism improvement, and relates to the field of remote sensing image classification. The method sequentially comprises the following steps of: a, acquiring a hyperspectral image data set, and artificially injecting random noise into image data; step b, constructing a neural network architecture search model for hyperspectral image crop classification, integrating a channel attention mechanism and a position scale perception mechanism so as to improve the calculation efficiency and the adaptability of the model, and enhancing the processing and anti-noise performance of difficult samples by using a focus loss function; c, searching a framework and performing framework evaluation; step d, performing model training by using the finally output architecture; and e, testing by using the complete test set, and outputting a classification result. According to the invention, by introducing the attention mechanism, the position and scale perception mechanism and the focus loss function based on the channel, the processing capability of the hyperspectral crop image mixed with noise is obviously enhanced, so that the classification precision and the robustness of the model are obviously improved.

The invention discloses a two-stage reinforcement learning SAT solving method for structure perception and a related device, and belongs to the field of chip design. Chip signal polarity and logic consistency information is introduced in a variable initialization stage, an initial feature vector for structure perception is constructed, and the limitation that traditional graph modeling can only represent a local logic relation is broken through; a two-stage reinforcement learning framework composed of a main path and a sampling path is designed, the sampling path is used for exploring a chip logic solution space, the main path is used for stably utilizing an optimal verification strategy, and the problems of unstable training convergence, single strategy and the like are effectively relieved. Based on a reward-guided feature updating mechanism, dynamic evolution of chip signal features and enhanced feedback are explicitly bound, the interpretability of a solver is improved, and the difference between neural reasoning and chip logic symbol reasoning is reduced. A signal assignment decision is executed based on GNN multi-round message passing results, and a verification track is continuously optimized, so that efficient chip design verification capability can be obtained without depending on an artificial heuristic mode.

The invention discloses a heuristic-based iterative legalization method and system for an analog circuit layout, and the method comprises the steps: dividing layout groups according to a special structure and an array module in a layout based on the initial layout of the layout containing overlapped devices or modules; constructing a constraint graph for describing the relative position relationship between the layout groups, and solving symmetric constraint conflicts in the layout so as to ensure that all symmetric constraints are met; performing simulation processing on a non-overlapping constraint, a minimum area constraint and a minimum spacing constraint through a code simulation solver to generate an initial legal layout; and iteratively adjusting the constraint graph based on a heuristic method, identifying a suspicious group and adjusting the constraint edge of the suspicious group, and re-verifying and repairing the symmetric constraint after each adjustment until the layout area converges. According to the method, the symmetric constraint is synchronously verified and repaired in each iterative optimization, so that the layout area is minimized on the premise of ensuring the layout legality, and the solving efficiency is greatly improved at the same time.

The present application relates to the technical field of GPU computing optimization, and in particular to a sparse attention computing method, device and medium for GPU, wherein the method realizes the high efficiency of long context reasoning through the geometric perception sparse attention framework of ball hashing, and combines a large-scale parallel hashing optimization algorithm and a load adaptive computing kernel. Compared with the existing sparse attention methods based on heuristics or gradient learning, the present application realizes higher retrieval recall rate, lower preprocessing overhead and efficient hardware adaptation to irregular sparse patterns.