Streaming RAG reduces latency via parallel tool queries

Arxiv CS.AI (top papers)

Streaming RAG issues tool queries in parallel with ongoing user input to reduce perceived latency. The paper characterizes tool-intent stabilization as a key factor determining when this approach provides benefit.

RAG

See 3 more sources

MCompassRAG: Topic Metadata as a Semantic Compass for Paragraph-Level Retrieval2 days agoAmirhossein Abaskohi, Raymond Li, Gaetano Cimino, Peter West, Giuseppe Carenini, Issam H. Laradji

Context Compression Is Not One Thing: Readable Symbolic Re-expression vs. Coherent Summary at Matched Budget4 days agoSisong Bei, Mikhail L. Arbuzov, Ziwei Dong, Dmitri Kalaev, Alexey Shvets

When Does Streaming Tool Use Help? Characterizing Tool-Intent Stabilization in Streaming Retrieval-Augmented Generation1 day agoElroy Galbraith

Arxiv CS.AI (top papers)·Xuedong HuZhiqing TangZhi YaoTian WangWeijia Jia

Jun 19, 4:00 AM