Continuous Learning LLM
A Knowledge Base which stores all the data shared by the users to the model without any deletions. Developing techniques to derive insights from the stored Knowledge Base. The current technique is to create a Knowledge Graph to organise the data. Knowledge Graph metadata is used to summarise data. Central Nodes in the graph indicate priority. As more data is received from the user, updates are made to the Knowledge Graph regularly. Vector Database is used for Retrieval Augmented Generation (RAG). Vector Database is updated with the data from the Knowledge Graph regularly. Superior RAG results because of: Highly organised input information being present in the Vector Database. High priority components from the data are clearly highlighted Summary of the entire data is present in the Vector Database.
Category tags:
For Innovators & Creators