{"slug": "order-task-conditioned-routing-for-retrieval-augmented-generation", "title": "ORDER: Task-Conditioned Routing for Retrieval-Augmented Generation", "summary": "Researchers introduced ORDER (Optimal Routing for Dynamic Evidence Retrieval), a query-conditioned Retrieval-Augmented Generation framework that jointly adapts indexing and retrieval to each incoming query, according to a new arXiv paper (2609.17012v1). ORDER discovers semantic clusters over a corpus's questions, learns a chunking strategy with metadata filtering and reranking per cluster, and routes queries at inference via nearest-centroid assignment, aided by a supervised query router (QRe) and a Uniform Multi-source Sampler (UMS) that splits the retrieval budget evenly across selected sources. Evaluated on large-scale heterogeneous historical archives, the framework consistently outperformed naive baselines and strong state-of-the-art RAG systems in complex expert-domain environments.", "body_md": "arXiv:2609.17012v1 Announce Type: new \nAbstract: Retrieval-Augmented Generation (RAG) pipelines typically rely on a fixed indexing and retrieval configuration determined at preprocessing time. This one-size-fits-all design is ill-suited to domain-expert settings, where heterogeneous queries require different chunking granularities, metadata constraints, and source-selection strategies. As a result, configurations that are effective for one family of queries often perform poorly for others. In this paper, we introduce ORDER (Optimal Routing for Dynamic Evidence Retrieval), a query-conditioned RAG framework that jointly adapts indexing and retrieval to the incoming query. Our approach first discovers semantic clusters over a given set of questions associated to a corpus and learns, for each cluster, a chunking strategy together with a suited metadata filtering and reranking configuration. At inference time, queries are routed to the appropriate pre-built index through nearest-centroid assignment. To further improve retrieval, we propose a supervised query router (QRe) that predicts which collections are most likely to contain relevant evidence, coupled with a Uniform Multi-source Sampler (UMS) that allocates the retrieval budget evenly across the selected sources. We evaluate our framework on large-scale, heterogeneous historical archives and show that conditioning both indexing and retrieval on the query consistently outperforms both naive baselines and strong state-of-the-art RAG systems in complex expert-domain environments.", "url": "https://wpnews.pro/news/order-task-conditioned-routing-for-retrieval-augmented-generation", "canonical_source": "https://www.machinebrief.com/news/order-task-conditioned-routing-for-retrieval-augmented-gener-rif5", "published_at": "2026-09-16 04:00:00+00:00", "updated_at": "2026-09-16 05:36:14.230274+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-research", "natural-language-processing", "large-language-models"], "entities": ["ORDER", "Optimal Routing for Dynamic Evidence Retrieval", "Retrieval-Augmented Generation", "QRe", "Uniform Multi-source Sampler", "arXiv"], "alternates": {"html": "https://wpnews.pro/news/order-task-conditioned-routing-for-retrieval-augmented-generation", "markdown": "https://wpnews.pro/news/order-task-conditioned-routing-for-retrieval-augmented-generation.md", "text": "https://wpnews.pro/news/order-task-conditioned-routing-for-retrieval-augmented-generation.txt", "jsonld": "https://wpnews.pro/news/order-task-conditioned-routing-for-retrieval-augmented-generation.jsonld"}}