Building natural language interface for Dataverse network based on Croissant ML standard

The new distributed network vision for AI is to create a shared data interface by querying multiple data nodes simultaneously. This distributed approach is based on the agile, community-driven standard called Croissant ML, allowing users to query and understand responses from various data platforms that support the standard, such as Dataverse, Kaggle, HuggingFace, and OpenML. The results are processed in a standardized way, converting metadata into a knowledge graph integrated with ontologies, and ingesting structured content into Large Language Models (LLMs) that act as reasoning engines, interfacing between humans and AI. The prototype enables "chatting" with individual data nodes in the network while considering ethical and privacy constraints, using only information shared through open metadata records.