aiqbench analysis
This paper introduces SRAT, an agentic reinforcement learning framework that dynamically decides when to retrieve external knowledge versus relying on parametric model knowledge for machine translation tasks. By optimizing retrieval actions based on downstream translation quality rather than simple semantic similarity, the method effectively handles partial-evidence regimes where external data supports but does not directly provide the target output. The results demonstrate significant gains over static retrieval baselines in idiom and knowledge-intensive translation scenarios.
Retrieve What Helps, Only When It Helps: Agentic Reinforcement Learning for Selective Retrieval-Augmented Machine Translation
arXiv:2610.19528v1 Announce Type: new Abstract: Retrieving external knowledge can improve large language model (LLM) translation by providing supplementary information such as idiom interpretations and terminology. However, existing retrieval-augmented machine translation systems typically apply predefined retrieval strategies uniformly across all inputs. We argue that translation operates in a partial-evidence regime: LLM's parametric knowledge suffices for many inputs, while external evidence benefits a subset and usually supports rather than directly provides the desired translation. This creates two challenges: deciding when to retrieve and what evidence to acquire. We introduce Selective Retrieval-Augmented Translation (SRAT), an agentic reinforcement learning framework that learns when to translate directly or retrieve, what to retrieve, and how to integrate these decisions by optimizing final translation quality. To align evidence retrieval with downstream translation quality rather than simple semantic similarity, we further develop joint Plackett--Luce optimization and translation-quality contrastive training for retriever-training. Experiments on idiom, knowledge-intensive, and document-level translation with offline or online retrieval show that SRAT consistently outperforms never- and always-retrieve baselines, and quality-aligned retriever training provides further gains when similarity-based retrieval is insufficient. The results demonstrate the benefit of treating retrieval as a learnable translation action rather than fixed preprocessing.
Abstract:Retrieving external knowledge can improve large language model (LLM) translation by providing supplementary information such as idiom interpretations and terminology. However, existing retrieval-augmented machine translation systems typically apply predefined retrieval strategies uniformly across all inputs. We argue that translation operates in a partial-evidence regime: LLM's parametric knowledge suffices for many inputs, while external evidence benefits a subset and usually supports rather than directly provides the desired translation. This creates two challenges: deciding when to retrieve and what evidence to acquire. We introduce Selective Retrieval-Augmented Translation (SRAT), an agentic reinforcement learning framework that learns when to translate directly or retrieve, what to retrieve, and how to integrate these decisions by optimizing final translation quality. To align evidence retrieval with downstream translation quality rather than simple semantic similarity, we further develop joint Plackett--Luce optimization and translation-quality contrastive training for retriever-training. Experiments on idiom, knowledge-intensive, and document-level translation with offline or online retrieval show that SRAT consistently outperforms never- and always-retrieve baselines, and quality-aligned retriever training provides further gains when similarity-based retrieval is insufficient. The results demonstrate the benefit of treating retrieval as a learnable translation action rather than fixed preprocessing.
Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
arXiv:2610.19528 [cs.CL]
(or arXiv:2610.19528v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2610.19528 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 289 words
Related coverage
Mastering Agentic Techniques: AI Agent Customization
NVIDIA · May 20, 2026
What's Continual Learning, and Why Might We Expect To See It In Advanced LLM Agents?
LessWrong · Jun 12, 2026
NVIDIA Vera CPU: Olympus Cores Built for Maximum Single-Thread Performance in Agentic AI
NVIDIA · Jul 21, 2026
[AINews] OpenAI DevDay 2026: Dots, 6.1 Sol, Ultrafast, Decisions API, Agents API, Spaces, Marketplace, and 1.2 Billion ChatGPT WAU
Latent Space · Sep 30, 2026