We asked an LLM agent a single question about Russian academia following the Russia–Ukraine conflict, and then iteratively issued follow-up queries on funding cuts, journal bias, and collaboration networks.
Instead of restarting each time, the agent progressively extended a shared analytical structure, linking four deep-dive analyses across 94 million relationships in the OpenAIRE knowledge graph.
This is ExBI: a new paradigm where LLM-driven exploration meets a hypergraph data model with unbiased sampling — growing with your curiosity, all on your local PC.
We asked an LLM agent a single question about Russian academia following the Russia–Ukraine conflict, and then iteratively issued follow-up queries on funding cuts, journal bias, and collaboration networks.
Instead of restarting each time, the agent progressively extended a shared analytical structure, linking four deep-dive analyses across 94 million relationships in the OpenAIRE knowledge graph.
This is ExBI: a new paradigm where LLM-driven exploration meets a hypergraph data model with unbiased sampling — growing with your curiosity, all on your local PC.
[dead]