Reason What Matters: Retrieval-Grounded Reasoning for Universal Multimodal Embeddings

• URL: http://arxiv.org/abs/2609.15296v1
• Date: Mon, 14 Sep 2026 09:48:05 GMT
• Title: Reason What Matters: Retrieval-Grounded Reasoning for Universal Multimodal Embeddings
• Abstract summary: Retrieval-aware Self-Distillation (RASD) constructs privileged guidance from input-supported evidence that distinguishes the positive item from retrieved hard negatives.<n>Retrieval-adaptive Inference (RAI) uses a retrieval confidence head to estimate the remaining retrieval utility of a partial CoT.<n>ReWAM achieves state-of-the-art retrieval performance while delivering up to 5x the inference throughput of competitive explicit-CoT UME methods.
• Score: 18.705254311226565
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Universal multimodal embedding (UME) learns unified representations across modalities, enabling a single model to support diverse retrieval tasks. Recent methods use Chain-of-Thought (CoT) reasoning to better interpret multimodal inputs before generating embeddings for complex retrieval tasks and further optimize this reasoning process through GRPO with retrieval-based rewards. However, two limitations hinder corpus-scale deployment. GRPO assigns all CoT tokens the same advantage, without identifying input-supported claims or evidence that distinguishes the positive from negatives. Moreover, generating a complete CoT before each embedding introduces substantial latency, even when a partial trace already provides sufficient retrieval evidence. To address these limitations, we propose Reason What Matters (ReWAM), a retrieval-grounded reasoning framework that uses retrieval feedback to guide both credit assignment and reasoning computation. Specifically, we introduce Retrieval-aware Self-Distillation (RASD), which constructs privileged guidance from input-supported evidence that distinguishes the positive item from retrieved hard negatives. An on-policy self-teacher uses this guidance to refine trajectory-level feedback into token-specific supervision for retrieval-relevant reasoning. We further develop Retrieval-adaptive Inference (RAI), which uses a retrieval confidence head to estimate the remaining retrieval utility of a partial CoT. It stops unproductive traces early and accelerates useful continuations with speculative decoding. Extensive experiments on MMEB-V2 and MRMR demonstrate that ReWAM achieves state-of-the-art retrieval performance while delivering up to 5x the inference throughput of competitive explicit-CoT UME methods. These results bridge the gap between retrieval quality and inference efficiency, making reasoning-enhanced UME practical for large-scale deployment.

Related papers

• Retrieval Grounding Latent Reasoning for Dense Retrieval [30.66182684102075]
  Reasoning-intensive retrieval requires reasoning needed to determine relevance under a given instruction.<n>We propose Retrieval Grounding Latent Reasoning (RGLT), a latent reasoning framework for dense retrieval.<n> Experiments on reasoning-intensive retrieval benchmarks show that RGLT consistently outperforms strong baselines.
  arXiv  Detail & Related papers  (2026-08-14T09:08:03Z)
• Generative Universal Multimodal Retrieval with Dual-role Identifiers [1.0981910170846492]
  We propose DrIG, a novel Generative framework for universal multimodal retrieval featuring Dual-role Identifiers.<n>DrIG consistently outperforms state-of-the-art generative multimodal baselines across diverse tasks, while hybrid reranking achieves a favorable efficiency-effectiveness trade-off against strong dense retrievers.
  arXiv  Detail & Related papers  (2026-08-13T09:10:09Z)
• Learning from Failures: Retrieval-Centric CoT via Hard Negatives for Unified Multimodal Retrieval [17.695551248974215]
  UniME-R1 is an embedder-adviser framework that learns to reason over initially retrieved candidates.<n>UniME-R1 consistently improves retrieval performance over strong baselines.
  arXiv  Detail & Related papers  (2026-08-06T14:04:56Z)
• WhisperRec: Latent Reasoning for Efficient Foundation Recommendation Models [51.88075626243057]
  We propose WhisperRec, an efficient latent reasoning framework for foundation recommendation models (FRMs)<n>WhisperRec compresses teacher-generated Chain-of-Thought (CoT) into learnable latent reasoning tokens.<n>We show that WhisperRec consistently outperforms explicit-CoT methods and conventional baselines.
  arXiv  Detail & Related papers  (2026-07-29T08:48:35Z)
• PLUME: Latent Reasoning Based Universal Multimodal Embedding [52.35354073629127]
  Universal multimodal embedding (UME) maps heterogeneous inputs into a shared retrieval space with a single model.<n>Recent approaches improve UME by generating explicit chain-of-thought (CoT) rationales before extracting embeddings.<n>We propose PLUME, a latent reasoning framework that advances UME by replacing verbalized CoT with a short autoregressive rollout of continuous latent states.
  arXiv  Detail & Related papers  (2026-04-02T14:04:53Z)
• Search-R2: Enhancing Search-Integrated Reasoning via Actor-Refiner Collaboration [49.9937230730202]
  We propose Search-R2, a novel Actor-Refiner collaboration framework that enhances reasoning through targeted intervention.<n>Our approach decomposes the generation process into an Actor, which produces initial reasoning trajectories.<n>We show that Search-R2 consistently outperforms strong RAG and RL-based baselines across model scales.
  arXiv  Detail & Related papers  (2026-02-03T15:32:09Z)
• Reinforced Efficient Reasoning via Semantically Diverse Exploration [73.41112984160992]
  Reinforcement learning with verifiable rewards (RLVR) has proven effective in enhancing the reasoning of large language models (LLMs)<n>We propose reinforced efficient reasoning via semantically diverse explorations, i.e., ROSE, for LLMs.<n>Our method incorporates a semantic-entropy-based branching strategy and an $varepsilon$-exploration mechanism.
  arXiv  Detail & Related papers  (2026-01-08T15:56:44Z)
• Adversarial Yet Cooperative: Multi-Perspective Reasoning in Retrieved-Augmented Language Models [72.4149653187766]
  We propose a Reasoner-Verifier framework named Adrialversa Reasoning RAG (ARR)<n>The Reasoner and Verifier engage in reasoning on retrieved evidence and critiquing each other's logic while being guided by process-aware advantage.<n> Experiments on multiple benchmarks demonstrate the effectiveness of our method.
  arXiv  Detail & Related papers  (2026-01-08T06:57:03Z)
• Thinking Forward and Backward: Multi-Objective Reinforcement Learning for Retrieval-Augmented Reasoning [137.33138614095435]
  Retrieval-augmented generation (RAG) has proven to be effective in mitigating hallucinations in large language models.<n>Recent efforts have incorporated search-based interactions into RAG, enabling iterative reasoning with real-time retrieval.<n>We propose Bi-RAR, a novel retrieval-augmented reasoning framework that evaluates each intermediate step jointly in both forward and backward directions.
  arXiv  Detail & Related papers  (2025-11-12T08:29:39Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.