Artificial intelligence · 2020
Retrieval-augmented generation
Give the model a search engine so it can cite documents instead of guessing.
Patrick Lewis and colleagues at Facebook AI combined a dense retriever with a generator so answers draw on fetched passages. RAG became the standard way to ground assistants in private documents and fresh facts.
What it built on 2
What it led to 1
Related measurements
The dashed line marks when it first worked.
Training compute of notable AI models Source: Epoch AI, Notable AI Models (CC BY 4.0)
ML hardware FLOP/s per dollar Source: Epoch AI, ML hardware (CC BY 4.0)