Other Databases Indexing TRIZ Cases
- Constrained Neural Network Training for Sparse, Explainable Models
Iteratively constrains neural-network weights and biases to remove non-conforming connections while preserving accuracy and improving explainability.
- Knowledge Graph Data Versioning with Immutable Published States
Status indicators separate editable drafts from immutable published versions, improving data integrity and auditability in knowledge graph systems.
- Hierarchical Document Graphs for Structure-Aware RAG Retrieval
Flat RAG retrieval can miss document hierarchy; context-embedded graph nodes preserve section relationships for relevant LLM responses.
- Universal Query Processing for Streaming and Batch Machine Data
A single query processor adapts to streaming or stored machine data, reducing system complexity while keeping diverse events searchable.
- DAG Node Sorting With Global Depth for AI Memory Efficiency
Conventional DAG sorting can create long edge spans and retain data across nodes; global-depth ordering improves locality and reduces hardware memory use.
- Personalized Knowledge Graphs from User Activity for AI Performance
To improve AI insights without a monolithic processor, the method segments user activity, forms episodes, and builds a personalized knowledge graph.
- Natural-Language Query Conversion with LLM Syntax Validation
An intent classifier and LLM generate cybersecurity queries, while lint feedback corrects syntax before data retrieval and visualization.
- Enterprise Knowledge Graphs for Accurate Generative AI Prompts
User-specific knowledge graphs add relevant enterprise relationships to LLM prompts, improving response accuracy without frequent retraining.
- Tree-Based Data Structure With Sorted and Unsorted Leaf Blocks
Separate sorted and unsorted leaf blocks let writers append new items quickly while readers use key order for efficient searches.
- Graph Data Structures Using Homomorphism to Reduce Redundancy
Graph homomorphisms collapse equivalent vertices into reduced connected components, limiting redundancy and simplifying entity resolution and data lineage.
- Metadata Graph Path Finding for Exponential Search Reduction
Traditional graph searches can examine exponentially many path permutations; metadata graphs filter irrelevant edge types and reduce traversal work.
- Internal Edge Selection for Protected Network Demand Routing
Clique and stream graphs narrow the internal-edge set for demand-satisfying paths, reducing links while preserving link-failure protection.
- Data Asset Access Governance Through Interconnection Graphs
A governance graph reveals direct and indirect data-set relationships, helping teams classify assets and support security and compliance.
- DAG-Orchestrated RAG Pipeline for Complex Query Reformulation
Complex or ambiguous queries move through DAG-based retrieval, reranking, refinement, and reformulation before grounded response generation.
- Self-Service Data Structure Discovery Through Automated Crawling
Manual inspection of repositories and batch contracts is replaced by crawled, indexed metadata and dashboard search for faster discovery.
- Event-Driven Graph Database Hydration from External Data
Event-driven hydration connects external data changes to schema-based rules, enabling near-real-time graph updates without manual loading.
- Graph Inference Models for Accurate Cross-Domain Resource Tagging
Ontology-enriched resource graphs and graph neural networks infer accurate tags across architectures, improving automated data-processing resource management.
- Manufacturing Process Graph Search With Structure and Table Comparison
Structure filtering followed by table-content comparison improves search reliability for complex manufacturing process graphs.
- Metric-Lens Graph Generation for Multidimensional Relationship Analysis
Metric-lens selection and entropy scoring generate topological graphs that expose important relationships in large datasets.