Retriever-Free Retrieval: Navigating Documents with Decision Models
arXiv cs.AI · 27 Oct 2026, 03:59 AEST
A new approach called DecisionRAG enhances document retrieval by using decision models instead of traditional retrieval methods. This method outperforms existing dense retrieval systems without requiring task-specific training.
Key facts
- While language models have become dramatically more capable, retrieval still largely relies on lexical matching or similarity in fixed embedding spaces.
- We propose DecisionRAG, which brings the capabilities of pretrained language models to retrieval using decision models that jointly evaluate candidate branches in a single forward pass, without generating any tokens.
- Without task-specific training or a first-stage retriever, DecisionRAG outperforms dense retrieval with an 8B reranker by 13.8 R@1 points on MultiDoc2Dial and 17.4 Hit@1 points on Natural Questions.
- Across four benchmarks, including a corpus of 2.68 million Wikipedia paragraphs, retrieval takes just 0.7-1.8 seconds per query.
Read aloud
A new approach called DecisionRAG enhances document retrieval by using decision models instead of traditional retrieval methods. This method outperforms existing dense retrieval systems without requiring task-specific training. While language models have become dramatically more capable, retrieval still largely relies on lexical matching or similarity in fixed embedding spaces. We propose DecisionRAG, which brings the capabilities of pretrained language models to retrieval using decision models that jointly evaluate candidate branches in a single forward pass, without generating any tokens. Without task-specific training or a first-stage retriever, DecisionRAG outperforms dense retrieval with an 8B reranker by thirteen point eight R@one points on MultiDoc2Dial and seventeen point four Hit@one points on Natural Questions. Across four benchmarks, including a corpus of two point six eight million Wikipedia paragraphs, retrieval takes just zero point seven-one point eight seconds per query.
Sources in this cluster
- Retriever-Free Retrieval: Navigating Documents with Decision Models
arXiv cs.AI · 27 Oct 2026, 03:59 AEST