MUMBAI, India, May 29 -- Intellectual Property India has published a patent application (202621044058 A) filed by Symbiosis International, Pune, Maharashtra, on April 6, for 'a multimodal knowledge graph-based system for contextual memory capture and retrieval using artificial intelligence.'
Inventor(s) include Sahil Goyal; Vaibhav Sharma; Girish Mahale; Vivek Hotchandani; Prof. Mayur Gaikwad; and Dr Sachin R Gaikwad.
The application for the patent was published on May 29, under issue no. 22/2026.
According to the abstract released by the Intellectual Property India: "A multimodal memory capture and retrieval system (100) combining real-time audio, video, image, and text processing with transformer-based AI models. The system comprises edge devices (101, 102, 103) capturing multimodal data transmitted through WebSocket streaming to a backend processing architecture. The backend transforms all modalities into unified textual descriptions (230) and generates structured subject-verb-object triplets (240) with metadata (241). These triplets are stored in a semantic knowledge graph (150) comprising nodes (310), edges (311), embeddings (320), and vector indices (321) representing past events and relationships. A hybrid retrieval engine (160) performs keyword search (341), embedding similarity (342), graph traversal (343), and temporal reasoning (344) to locate relevant memory units. Retrieved results are processed by a retrieval-augmented generation module (170) generating contextually appropriate natural-language responses (370). The invention enables privacy-preserving, offline, real-time multimodal memory reconstruction for personal knowledge management, assistive cognition, and intelligent contextual systems."
Disclaimer: Curated by HT Syndication.
