Canon is a research project that I’ve started to work on related to Retrieval Augmented Generation and AI. The goal of the project is to create a semantic knowledge base and a framework for retrieval augmented generation that can be used at interference time and during training in order to improve grounding in large language models.