Apple researchers train learned networks to speed up RAG retrieval
AIApple researchers present SupportNet and KeyNet, small neural networks trained offline to route queries to their best-matching keys faster than centroid-based search, at ICML 2026. The paper, Amortizing Maximum Inner Product Search with Learned Support Functions, says the approach trades an upfront training cost for faster, more accurate retrieval at inference time.
Source details
Apple Machine Learning Research · machinelearning.apple.comPublished · added here
Original title: Faster Retrieval for Better, More Efficient RAG