Entropy-Gradient Grounding: Training-Free Evidence Retrieval in Vision-Language Models

• URL: http://arxiv.org/abs/2604.08456v1
• Date: Thu, 09 Apr 2026 16:51:42 GMT
• Title: Entropy-Gradient Grounding: Training-Free Evidence Retrieval in Vision-Language Models
• Abstract summary: We propose a training-free, model-intrinsic grounding method for vision-language models.<n>We compute the entropy of the model's next-token distribution and backpropagate it to the visual token embeddings to obtain an entropy-gradient relevance map.<n>We then extract and rank multiple coherent regions to support multi-evidence queries, and introduce an iterative zoom-and-reground procedure with a spatial-entropy stopping rule.
• Score: 77.3748853516374
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Despite rapid progress, pretrained vision-language models still struggle when answers depend on tiny visual details or on combining clues spread across multiple regions, as in documents and compositional queries. We address this by framing grounding as test-time evidence retrieval: given a query, the model should actively identify where to look next to resolve ambiguity. To this end, we propose a training-free, model-intrinsic grounding method that uses uncertainty as supervision. Specifically, we compute the entropy of the model's next-token distribution and backpropagate it to the visual token embeddings to obtain an entropy-gradient relevance map, without auxiliary detectors or attention-map heuristics. We then extract and rank multiple coherent regions to support multi-evidence queries, and introduce an iterative zoom-and-reground procedure with a spatial-entropy stopping rule to avoid over-refinement. Experiments on seven benchmarks across four VLM architectures demonstrate consistent improvements over existing methods, with the largest gains on detail-critical and high-resolution settings, while also producing more interpretable evidence localizations.

Related papers

• From Uncertainty to Determinism: Coarse-to-Fine Visual Floorplan Localization without Ray Matching [12.273420964874058]
  We propose a coarse-to-fine visual FLoc framework that progresses from uncertainty to determinism.<n>In the refinement stage, we propose a localized refiner that predicts bounded sub-meter pose residuals from candidate-centered crops.<n>Our method effectively balances global multi-hypothesis tracking and local sub-meter refinement without requiring any offline map preprocessing or test-time lookup tables.
  arXiv  Detail & Related papers  (2026-07-29T12:10:00Z)
• REVERSE: Reinforcing Evidence Verification and Search for Agentic Image geo-localization [15.44857530497746]
  We present REVERSE, a framework that teaches three intermediate decisions: where to look, what to query, and what evidence to trust.<n>An offline search cache makes retrieval observations stable and reusable during reinforcement learning, enabling dense supervision over noisy search results.<n>With a 4B model, REVERSE outperforms strong retrieval-augmented baselines and rivals substantially larger models on Im2GPS3k and YFCC4k.
  arXiv  Detail & Related papers  (2026-05-26T11:20:09Z)
• SpatialReward: Verifiable Spatial Reward Modeling for Fine-Grained Spatial Consistency in Text-to-Image Generation [62.55421542903781]
  We present textbfSpatialReward, a verifiable reward model explicitly designed to evaluate spatial layouts in generated images.<n>Experiments on Stable Diffusion and FLUX show that incorporating SpatialReward into RL training consistently improves spatial consistency and overall generation quality.
  arXiv  Detail & Related papers  (2026-03-23T17:26:35Z)
• VIEW2SPACE: Studying Multi-View Visual Reasoning from Sparse Observations [47.94531550391802]
  Multi-view visual reasoning is essential for intelligent systems that must understand complex environments from sparse and discrete viewpoints.<n>In real-world scenarios, reasoning across views requires integrating partial observations without explicit guidance.<n>We leverage physically grounded simulation to construct diverse, high-fidelity 3D scenes with precise per-view metadata.
  arXiv  Detail & Related papers  (2026-03-17T13:36:30Z)
• Multi-session Localization and Mapping Exploiting Topological Information [1.3316173544577008]
  We propose a novel multi-session framework that builds on map-based localization.<n>Our approach incorporates a topology-informed, uncertainty-aware decision-making mechanism.<n>We validate our method on sequences overlapping from datasets and demonstrate its effectiveness in a real-world mine-like environment.
  arXiv  Detail & Related papers  (2026-02-19T10:17:46Z)
• DeepImageSearch: Benchmarking Multimodal Agents for Context-Aware Image Retrieval in Visual Histories [52.57197752244638]
  We introduce DeepImageSearch, a novel agentic paradigm that reformulates image retrieval as an autonomous exploration task.<n>Models must plan and perform multi-step reasoning over raw visual histories to locate targets based on implicit contextual cues.<n>We construct DISBench, a challenging benchmark built on interconnected visual data.
  arXiv  Detail & Related papers  (2026-02-11T12:51:10Z)
• WeMMU: Enhanced Bridging of Vision-Language Models and Diffusion Models via Noisy Query Tokens [69.97021957331326]
  We propose Noisy Query Tokens, which learn a distributed representation space between the VLM and Diffusion Model via end-to-end optimization.<n>We also introduce a VAE branch with linear projection to recover fine-grained image details.
  arXiv  Detail & Related papers  (2025-12-02T09:02:20Z)
• Towards General Visual-Linguistic Face Forgery Detection [95.73987327101143]
  Deepfakes are realistic face manipulations that can pose serious threats to security, privacy, and trust. Existing methods mostly treat this task as binary classification, which uses digital labels or mask signals to train the detection model. We propose a novel paradigm named Visual-Linguistic Face Forgery Detection(VLFFD), which uses fine-grained sentence-level prompts as the annotation.
  arXiv  Detail & Related papers  (2023-07-31T10:22:33Z)
• Far Away in the Deep Space: Dense Nearest-Neighbor-Based Out-of-Distribution Detection [33.78080060234557]
  Nearest-Neighbors approaches have been shown to work well in object-centric data domains. We show that nearest-neighbor approaches also yield state-of-the-art results on dense novelty detection in complex driving scenes.
  arXiv  Detail & Related papers  (2022-11-12T13:32:19Z)
• Bayesian Graph Contrastive Learning [55.36652660268726]
  We propose a novel perspective of graph contrastive learning methods showing random augmentations leads to encoders. Our proposed method represents each node by a distribution in the latent space in contrast to existing techniques which embed each node to a deterministic vector. We show a considerable improvement in performance compared to existing state-of-the-art methods on several benchmark datasets.
  arXiv  Detail & Related papers  (2021-12-15T01:45:32Z)
• Glimpse-Attend-and-Explore: Self-Attention for Active Visual Exploration [47.01485765231528]
  Active visual exploration aims to assist an agent with a limited field of view to understand its environment based on partial observations. We propose the Glimpse-Attend-and-Explore model which employs self-attention to guide the visual exploration instead of task-specific uncertainty maps. Our model provides encouraging results while being less dependent on dataset bias in driving the exploration.
  arXiv  Detail & Related papers  (2021-08-26T11:41:03Z)

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.