Scaling Representation Diversity: Modulated Attention and Reconstructive Regularization for Visual Grounding

• URL: http://arxiv.org/abs/2608.12748v1
• Date: Thu, 13 Aug 2026 02:57:21 GMT
• Title: Scaling Representation Diversity: Modulated Attention and Reconstructive Regularization for Visual Grounding
• Abstract summary: Referring Expression (REC) is commonly studied under dataset-specific fine-tuning.<n>In this work, we revisit REC from the perspective of unified open-vocabulary grounding.<n>To preserve representation diversity, we propose a holistic data-model co-design framework.
• Score: 13.310120232835722
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Referring Expression Comprehension (REC) is commonly studied under dataset-specific fine-tuning, resulting in specialist models with limited cross-dataset generalization. In this work, we revisit REC from the perspective of unified open-vocabulary grounding and identify representation degeneration as a key obstacle to scaling a single generalist model. To preserve representation diversity, we propose a holistic data-model co-design framework. Architecturally, we introduce the Modulated Attention-Contrastive Head (mACH) for efficient token-level vision-language alignment and a text-conditioned JEPA auxiliary stream that provides complementary gradient support to preserve alignment-active representations without inference overhead. On the data side, we introduce Objects365-Caption, enriching Objects365 with context-aware referring expressions for large-scale language supervision. We further provide a theoretical analysis showing that complementary gradient subspaces preserve alignment capacity and thereby scale representation diversity. Extensive experiments demonstrate that our single-checkpoint framework achieves highly competitive performance on standard REC benchmarks while exhibiting strong generalization across heterogeneous grounding datasets without benchmark-specific adaptation.

Related papers

• Generative Archetype-Grounded Item Representations for Sequential Recommendation [45.91194797303667]
  GenAIR is a general framework that empowers sequential recommendation with Generative Archetype-grounded Item Representations.<n>We introduce a behavioral calibration objective, which explicitly incorporates behavioral signals from actual interactions.<n>Experiments conducted on three real-world datasets demonstrate that GenAIR significantly improves the performance of various sequential recommendation models.
  arXiv  Detail & Related papers  (2026-06-09T15:59:14Z)
• Beyond Point-Wise Matching: Structural Representation Alignment for Accelerating Diffusion Transformers [93.3976834364707]
  We propose sREPA, a structural REPresentation Alignment framework to enforce consistency in the relational geometry of feature maps.<n>By encouraging the model to internalize holistic spatial layouts and structural correlations from pre-trained features, sREPA achieves faster and more stable convergence.
  arXiv  Detail & Related papers  (2026-05-16T12:01:04Z)
• Bridging Visual Representation and Reinforcement Learning from Verifiable Rewards in Large Vision-Language Models [35.97638821723309]
  KAWHI is a plug-and-play reward reweighting mechanism that explicitly incorporates structured visual information into uniform reward policy optimization methods.<n>It adaptively localizes semantically salient regions through hierarchical geometric aggregation, identifies vision-critical attention heads via structured attribution, and performs paragraph-level credit reallocation to align spatial visual evidence with semantically decisive reasoning steps.
  arXiv  Detail & Related papers  (2026-03-28T18:40:14Z)
• Beyond the Academic Monoculture: A Unified Framework and Industrial Perspective for Attributed Graph Clustering [75.50670592447219]
  Attributed Graph Clustering (AGC) is a fundamental unsupervised task that partitions nodes into cohesive groups by jointly modeling structural topology and node attributes.<n>This survey provides a comprehensive, industrially grounded review of AGC from three complementary perspectives.
  arXiv  Detail & Related papers  (2026-03-21T14:15:34Z)
• Asymmetric Hierarchical Anchoring for Audio-Visual Joint Representation: Resolving Information Allocation Ambiguity for Robust Cross-Modal Generalization [19.721857318111734]
  We propose Asymmetric Hierarchical Anchoring (AHA) to enforce directional information allocation.<n>We replace fragile mutual information estimators with a GRL-based adversarial decoupler that explicitly suppresses semantic leakage.<n>AHA consistently outperforms symmetric baselines in cross-modal transfer.
  arXiv  Detail & Related papers  (2026-02-03T14:14:03Z)
• CoT Referring: Improving Referring Expression Tasks with Grounded Reasoning [67.18702329644526]
  CoT Referring enhances model reasoning across modalities through a structured, chain-of-thought training data structure.<n>We restructure the training data to enforce a new output form, providing new annotations for existing datasets.<n>We also integrate detection and segmentation capabilities into a unified MLLM framework, training it with a novel adaptive weighted loss to optimize performance.
  arXiv  Detail & Related papers  (2025-10-03T08:50:21Z)
• CRIA: A Cross-View Interaction and Instance-Adapted Pre-training Framework for Generalizable EEG Representations [52.251569042852815]
  CRIA is an adaptive framework that utilizes variable-length and variable-channel coding to achieve a unified representation of EEG data across different datasets.<n>The model employs a cross-attention mechanism to fuse temporal, spectral, and spatial features effectively.<n> Experimental results on the Temple University EEG corpus and the CHB-MIT dataset show that CRIA outperforms existing methods with the same pre-training conditions.
  arXiv  Detail & Related papers  (2025-06-19T06:31:08Z)
• Entity-enhanced Adaptive Reconstruction Network for Weakly Supervised Referring Expression Grounding [214.8003571700285]
  Weakly supervised Referring Expression Grounding (REG) aims to ground a particular target in an image described by a language expression. We design an entity-enhanced adaptive reconstruction network (EARN) EARN includes three modules: entity enhancement, adaptive grounding, and collaborative reconstruction.
  arXiv  Detail & Related papers  (2022-07-18T05:30:45Z)
• Cluster-level Feature Alignment for Person Re-identification [16.01713931617725]
  This paper probes another feature alignment modality, namely cluster-level feature alignment across whole dataset. We propose anchor loss and investigate many variants of cluster-level feature alignment, which consists of iterative aggregation and alignment from overview of dataset.
  arXiv  Detail & Related papers  (2020-08-15T23:47:47Z)
• Target-Embedding Autoencoders for Supervised Representation Learning [111.07204912245841]
  This paper analyzes a framework for improving generalization in a purely supervised setting, where the target space is high-dimensional. We motivate and formalize the general framework of target-embedding autoencoders (TEA) for supervised prediction, learning intermediate latent representations jointly optimized to be both predictable from features as well as predictive of targets.
  arXiv  Detail & Related papers  (2020-01-23T02:37:10Z)

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.