Intelligent Continuous Learning Knowledge Graph

• The project seeks to create an Intelligent Continuous Learning Knowledge Graph to improve domain knowledge using Agentic AI and Multi-Modal Knowledge Extraction (MMKE).
• The project falls under the stream of Data Intelligence and Information Retrieval.

Problem Statement

• Traditional search engines rely on keyword-based retrieval, leading to irrelevant results.
• Existing search engines lack contextual understanding.
• Traditional search engines use static knowledge, which does not evolve dynamically with user intent.

Objective

• The project aims to design a continuous learning knowledge graph that retrieves research papers, YouTube videos, textbooks, GitHub repositories, and Stack Overflow discussions based on semantic search.
• The project seeks to integrate Multi-Modal Knowledge Extraction (MMKE) to process and link knowledge across text, video, and code.
• NLP techniques will be applied to improve search queries and results ranking.
• Designing an intuitive user interface for easy interaction and visualization of search results is aimed for.
• The project will develop an autonomous research assistant using Agentic AI to refine queries and improve search relevance.

Scope

• The target users are researchers, students, and developers seeking domain-specific resources.
• The project will have a web-based interface for input queries.
• Multi-Modal Knowledge Extraction (MMKE) will be used for text, videos, and code.
• NLP will drive semantic understanding of queries.
• A Continuous Learning knowledge graph will perform structured result organization.
• The project aims to retrieve research papers via APIs like Semantic Scholar and ArXiv
• Will retrieve YouTube videos, textbooks (Open Library), and GitHub repositories along with Stack Overflow discussions.
• There will be integration of multiple AI agents for query understanding, search ranking, and result summarization.
• Evaluation metrics will measure precision, recall, and relevance of search results.

Technology to be used

• The hardware to be used is Intel Core i7 (12th Gen) / AMD Ryzen 7, with 16/32GB RAM and 1TB SSD.

Software

• Frontend: React.js
• Backend: Node.js, Express.js
• Database: MongoDB (for metadata storage), Neo4j (for knowledge graph).
• NLP Models: BERT, OpenAI embeddings, spaCy.
• APIs: Semantic Scholar API, YouTube API, GitHub API, Open Library API.
• Agentic AI Frameworks: LangChain, AutoGPT, BabyAGI.
• LLM-based Processing: OpenAI's GPT APIs, Hugging Face models.

Societal Contributions

• Enhancing research accessibility will help students, researchers, and developers find relevant papers, videos, and repositories quickly.
• Reducing information overload by filtering out irrelevant search results using semantic understanding.
• Improving educational outcomes by suggesting high-quality study materials based on query intent.
• Bridging the gap between theory and implementation by linking academic research with practical GitHub implementations and Stack Overflow discussions.