POINTS-Long: Adaptive Dual-Mode Visual Reasoning in MLLMs

• URL: http://arxiv.org/abs/2604.11627v1
• Date: Mon, 13 Apr 2026 15:38:22 GMT
• Title: POINTS-Long: Adaptive Dual-Mode Visual Reasoning in MLLMs
• Abstract summary: POINTS-Long is a native dual-mode MLLM featuring dynamic visual token scaling inspired by the human visual system.<n>Our work provides new insights into the design of future MLLMs and lays the foundation for adaptive and efficient long-form visual understanding.
• Score: 64.72502363697518
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Multimodal Large Language Models (MLLMs) have recently demonstrated remarkable capabilities in cross-modal understanding and generation. However, the rapid growth of visual token sequences--especially in long-video and streaming scenarios--poses a major challenge to their scalability and real-world deployment. Thus, we introduce POINTS-Long, a native dual-mode MLLM featuring dynamic visual token scaling inspired by the human visual system. The model supports two complementary perception modes: focus mode and standby mode, enabling users to dynamically trade off efficiency and accuracy during inference. On fine-grained visual tasks, the focus mode retains the optimal performance, while on long-form general visual understanding, the standby mode retains 97.7-99.7% of the original accuracy using only 1/40-1/10th of the visual tokens. Moreover, POINTS-Long natively supports streaming visual understanding via a dynamically detachable KV-cache design, allowing efficient maintenance of ultra-long visual memory. Our work provides new insights into the design of future MLLMs and lays the foundation for adaptive and efficient long-form visual understanding.

Related papers

• Vision Inference Former: Sustaining Visual Consistency in Multimodal Large Language Models [58.9387276447485]
  Vision Inference Former (VIF) is a lightweight architectural module that establishes a bridge between pure visual representations and the model's output space.<n>We conduct experiments on 14 benchmark tasks covering general reasoning, OCR, table understanding, vision-centric evaluation, and hallucination.
  arXiv  Detail & Related papers  (2026-05-18T10:04:22Z)
• Predictive Regularization Against Visual Representation Degradation in Multimodal Large Language Models [84.94288033791346]
  We conduct a detailed diagnostic analysis to unveil a pervasive issue: visual representation degradation in MLLMs.<n>We attribute this phenomenon to a visual sacrifice driven by the singular text-generation objective, where the model compromises its visual fidelity to optimize for answer generation.<n>We propose Predictive Regularization to force degraded intermediate features to predict initial visual features, thereby maintaining the inherent visual attributes of the MLLM's internal representations.
  arXiv  Detail & Related papers  (2026-03-21T13:10:37Z)
• Blink: Dynamic Visual Token Resolution for Enhanced Multimodal Understanding [43.63398524449102]
  Humans perceive complex scenes efficiently by dynamically scanning and focusing on salient regions in a sequential "blink-like" process.<n>We propose Blink, a dynamic visual token resolution framework that emulates the human-inspired process within a single forward pass.<n>Blink balances broad exploration and fine-grained focus, thereby enhancing visual perception adaptively and efficiently.
  arXiv  Detail & Related papers  (2025-12-11T11:27:25Z)
• Unleashing the Intrinsic Visual Representation Capability of Multimodal Large Language Models [58.91911788912665]
  We propose Latent Visual Reconstruction (LaVer), a novel training framework that facilitates MLLMs in learning more discrimi visual representations.<n>Our method offers direct visual activation to MLLMs, which exhibit increased visual attention allocation, indicating enhanced utilization of visual information.
  arXiv  Detail & Related papers  (2025-12-06T04:20:13Z)
• DINOv2 Driven Gait Representation Learning for Video-Based Visible-Infrared Person Re-identification [30.593882551803855]
  Video-based Visible-Infrared person re-identification (VVI-ID) aims to retrieve the same pedestrian across visible and infrared modalities from modality.<n>To address these challenges, we propose a Gait Representation Learning framework that leverages the rich visual priors of DINOv2 to learn gait features complementary to appearance cues.<n>Specifically, we introduce a Semantic-Aware Silhouette and Gait Learning (GL) model, which generates and enhances silhouette representations with general-purpose semantic priors from DINOv2.
  arXiv  Detail & Related papers  (2025-11-06T11:21:13Z)
• Lifting the Veil on Visual Information Flow in MLLMs: Unlocking Pathways to Faster Inference [28.24397677839652]
  Multimodal large language models (MLLMs) improve performance on vision-language tasks by integrating visual features from pre-trained vision encoders into large language models.<n>How MLLMs process and utilize visual information remains unclear.<n>We propose Hierarchical Modality-Aware Pruning (HiMAP), a plug-and-play inference acceleration method that dynamically prunes image tokens at specific layers, reducing computational costs by approximately 65% without sacrificing performance.
  arXiv  Detail & Related papers  (2025-03-17T12:31:23Z)
• MMFuser: Multimodal Multi-Layer Feature Fuser for Fine-Grained Vision-Language Understanding [39.68348330596116]
  We propose modelname, a simple yet effective multi-layer feature fuser that efficiently integrates deep and shallow features from Vision Transformers (ViTs) Specifically, it leverages semantically aligned deep features as queries to dynamically extract missing details from shallow features. modelnameachieves significant improvements in visual representation and benchmark performance.
  arXiv  Detail & Related papers  (2024-10-15T17:55:22Z)
• Chain-of-Spot: Interactive Reasoning Improves Large Vision-Language Models [81.71651422951074]
  Chain-of-Spot (CoS) method is a novel approach that enhances feature extraction by focusing on key regions of interest. This technique allows LVLMs to access more detailed visual information without altering the original image resolution. Our empirical findings demonstrate a significant improvement in LVLMs' ability to understand and reason about visual content.
  arXiv  Detail & Related papers  (2024-03-19T17:59:52Z)
• Harnessing Diffusion Models for Visual Perception with Meta Prompts [68.78938846041767]
  We propose a simple yet effective scheme to harness a diffusion model for visual perception tasks. We introduce learnable embeddings (meta prompts) to the pre-trained diffusion models to extract proper features for perception. Our approach achieves new performance records in depth estimation tasks on NYU depth V2 and KITTI, and in semantic segmentation task on CityScapes.
  arXiv  Detail & Related papers  (2023-12-22T14:40:55Z)

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.