Divide-and-Conquer Inference for Large-Scale Visual Recognition with Multimodal Large Language Models

• URL: http://arxiv.org/abs/2605.24799v1
• Date: Sun, 24 May 2026 01:07:05 GMT
• Title: Divide-and-Conquer Inference for Large-Scale Visual Recognition with Multimodal Large Language Models
• Abstract summary: Divide-and-Conquer Inference (DCI) is a novel test-time scaling strategy for visual recognition with MLLMs.<n>DCI decomposes complex global classification tasks into simpler, localized subproblems and employs a dynamic pruning mechanism to compress the search space.<n>As a model-agnostic, plug-and-play paradigm, DCI offers an efficient approach for scaling the inferential precision of MLLMs in large-scale scenarios.
• Score: 12.459927405623624
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Multimodal Large Language Models (MLLMs) have demonstrated strong capabilities across a wide range of vision language tasks. However, when applied to large scale image classification, their performance degrades significantly as the label space expands a phenomenon we define as Performance Collapse in Long Sequence Recognition. Through an information theoretic analysis, we reveal that this collapse stems from a fundamental conflict between the escalating information entropy and the prominent attention dilution and decay within attention mechanisms, which impairs the model's ability to maintain a sufficient signal-to-noise ratio when processing extremely long prompts. To mitigate this, we propose Divide-and-Conquer Inference (DCI), a novel test-time scaling strategy for visual recognition with MLLMs. DCI recursively decomposes complex global classification tasks into multiple simpler, localized subproblems and employs a dynamic pruning mechanism to compress the search space. This method effectively improves the local signal to noise ratio and model accuracy by mitigating the inherent weight dilution issues in long-sequence inference. Moreover, while traditional self-attention incurs a prohibitive quadratic computational complexity, DCI achieves more favorable scaling behavior and substantially accelerates inference in large scale classification scenarios. Extensive experiments on benchmarks such as ImageNet-1K and ImageNet-21K demonstrate that DCI consistently improves classification accuracy. This enables lightweight open-source models to rival or even surpass frontier closed-source giants without any additional training or fine-tuning. As a model-agnostic, plug-and-play paradigm, DCI offers an efficient approach for scaling the inferential precision of MLLMs in large-scale scenarios.

Related papers

• Focus When Necessary: Adaptive Routing and Collaborative Grounding for Training-Free Visual Grounding [56.38477453373368]
  LazyMCoT is a dynamic and training-free framework that adaptively allocates visual grounding efforts based on sample difficulty.<n>LazyMCoT rivals training-based approaches by simultaneously improving reasoning accuracy and reducing average inference latency.
  arXiv  Detail & Related papers  (2026-06-15T03:17:44Z)
• Max-pooling Network Revisited: Analyzing the Role of Semantic Probability in Multiple Instance Learning for Hallucination Detection [31.935824861650005]
  Hallucination detection has become increasingly important for improving the reliability of large language models.<n>Recently, hybrid approaches such as HaMI, which combine semantic consistency with internal model states via Multiple Instance Learning (MIL), have achieved state-of-the-art performance.<n>In this work, we first provide a theoretical analysis of HaMI in terms of decision margins, revealing that scaling internal states with semantic consistency leads to an enlarged decision margin.
  arXiv  Detail & Related papers  (2026-05-09T10:30:30Z)
• WSD-MIL: Window Scale Decay Multiple Instance Learning for Whole Slide Image Classification [2.760935655675299]
  Window scale decay MIL (WSD-MIL) is designed to enhance the capacity to model tumor regions of varying scales.<n>WSD-MIL achieves state-of-the-art performance on the CAMELYON16 and TCGA-BRCA datasets while reducing 62% of the computational memory.
  arXiv  Detail & Related papers  (2025-12-23T02:10:24Z)
• Multiscale Aggregated Hierarchical Attention (MAHA): A Game Theoretic and Optimization Driven Approach to Efficient Contextual Modeling in Large Language Models [0.0]
  Multiscale Aggregated Hierarchical Attention (MAHA) is a novel architectural framework that reformulates the attention mechanism through hierarchical decomposition and mathematically rigorous aggregation.<n>MAHA dynamically partitions the input sequence into hierarchical scales via learnable downsampling operators.<n> Experimental evaluations demonstrate that MAHA achieves superior scalability; empirical FLOPs analysis confirms an 81% reduction in computational cost at a sequence length of 4096 compared to standard attention.
  arXiv  Detail & Related papers  (2025-12-16T21:27:21Z)
• FAIM: Frequency-Aware Interactive Mamba for Time Series Classification [87.84511960413715]
  Time series classification (TSC) is crucial in numerous real-world applications, such as environmental monitoring, medical diagnosis, and posture recognition.<n>We propose FAIM, a lightweight Frequency-Aware Interactive Mamba model.<n>We show that FAIM consistently outperforms existing state-of-the-art (SOTA) methods, achieving a superior trade-off between accuracy and efficiency.
  arXiv  Detail & Related papers  (2025-11-26T08:36:33Z)
• ScaleGNN: Towards Scalable Graph Neural Networks via Adaptive High-order Neighboring Feature Fusion [73.85920403511706]
  We propose ScaleGNN, a novel framework that adaptively fuses multi-hop node features for scalable and effective graph learning.<n>We show that ScaleGNN consistently outperforms state-of-the-art GNNs in both predictive accuracy and computational efficiency.
  arXiv  Detail & Related papers  (2025-04-22T14:05:11Z)
• Exploring Training and Inference Scaling Laws in Generative Retrieval [50.82554729023865]
  Generative retrieval reformulates retrieval as an autoregressive generation task, where large language models generate target documents directly from a query.<n>We systematically investigate training and inference scaling laws in generative retrieval, exploring how model size, training data scale, and inference-time compute jointly influence performance.
  arXiv  Detail & Related papers  (2025-03-24T17:59:03Z)
• USDRL: Unified Skeleton-Based Dense Representation Learning with Multi-Grained Feature Decorrelation [24.90512145836643]
  We introduce a Unified Skeleton-based Dense Representation Learning framework based on feature decorrelation.<n>We show that our approach significantly outperforms the current state-of-the-art (SOTA) approaches.
  arXiv  Detail & Related papers  (2024-12-12T12:20:27Z)
• Efficient High-Resolution Visual Representation Learning with State Space Model for Human Pose Estimation [60.80423207808076]
  Capturing long-range dependencies while preserving high-resolution visual representations is crucial for dense prediction tasks such as human pose estimation.<n>We propose the Dynamic Visual State Space (DVSS) block, which augments visual state space models with multi-scale convolutional operations.<n>We build HRVMamba, a novel model for efficient high-resolution representation learning.
  arXiv  Detail & Related papers  (2024-10-04T06:19:29Z)
• Adversarial Feature Augmentation and Normalization for Visual Recognition [109.6834687220478]
  Recent advances in computer vision take advantage of adversarial data augmentation to ameliorate the generalization ability of classification models. Here, we present an effective and efficient alternative that advocates adversarial augmentation on intermediate feature embeddings. We validate the proposed approach across diverse visual recognition tasks with representative backbone networks.
  arXiv  Detail & Related papers  (2021-03-22T20:36:34Z)
• Crowd Counting via Hierarchical Scale Recalibration Network [61.09833400167511]
  We propose a novel Hierarchical Scale Recalibration Network (HSRNet) to tackle the task of crowd counting. HSRNet models rich contextual dependencies and recalibrating multiple scale-associated information. Our approach can ignore various noises selectively and focus on appropriate crowd scales automatically.
  arXiv  Detail & Related papers  (2020-03-07T10:06:47Z)

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.