Cover Image

Downscaling Intelligence: Exploring Perception and Reasoning Bottlenecks in Small Multimodal Models

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Endo, Mark, Yeung-Levy, Serena
Format:	Preprint
Published:	2025
Subjects:	Computer Vision and Pattern Recognition
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Feather the Throttle: Revisiting Visual Token Pruning for Vision-Language Model Acceleration
by: Endo, Mark, et al.
Published: (2024)

Continuous Perception Matters: Diagnosing Temporal Integration Failures in Multimodal Models
by: Wang, Zeyu, et al.
Published: (2024)

Foundation Models Secretly Understand Neural Network Weights: Enhancing Hypernetwork Architectures with Foundation Models
by: Gu, Jeffrey, et al.
Published: (2025)

Systematic Evaluation of Large Vision-Language Models for Surgical Artificial Intelligence
by: Rau, Anita, et al.
Published: (2025)

Zero-shot Action Localization via the Confidence of Large Vision-Language Models
by: Aklilu, Josiah, et al.
Published: (2024)

Do VLMs Perceive or Recall? Probing Visual Perception vs. Memory with Classic Visual Illusions
by: Sun, Xiaoxiao, et al.
Published: (2026)

The Impact of Image Resolution on Biomedical Multimodal Large Language Models
by: Chen, Liangyu, et al.
Published: (2025)

Seeing Is Believing? A Benchmark for Multimodal Large Language Models on Visual Illusions and Anomalies
by: Hou, Wenjin, et al.
Published: (2026)

Multi-Human Mesh Recovery with Transformers
by: Wang, Zeyu, et al.
Published: (2024)

SciVideoBench: Benchmarking Scientific Video Reasoning in Large Multimodal Models
by: Deng, Andong, et al.
Published: (2025)

Diffusion-HPC: Synthetic Data Generation for Human Mesh Recovery in Challenging Domains
by: Weng, Zhenzhen, et al.
Published: (2023)

DeforHMR: Vision Transformer with Deformable Cross-Attention for 3D Human Mesh Recovery
by: Heo, Jaewoo, et al.
Published: (2024)

Just Shift It: Test-Time Prototype Shifting for Zero-Shot Generalization with Vision-Language Models
by: Sui, Elaine, et al.
Published: (2024)

Motion Diffusion-Guided 3D Global HMR from a Dynamic Camera
by: Heo, Jaewoo, et al.
Published: (2024)

Viewpoint Textual Inversion: Discovering Scene Representations and 3D View Control in 2D Diffusion Models
by: Burgess, James, et al.
Published: (2023)

VisuRiddles: Fine-grained Perception is a Primary Bottleneck for Multimodal Large Language Models in Abstract Visual Reasoning
by: Yan, Hao, et al.
Published: (2025)

Revisiting Active Learning in the Era of Vision Foundation Models
by: Gupte, Sanket Rajan, et al.
Published: (2024)

Breaking Dual Bottlenecks: Evolving Unified Multimodal Models into Self-Adaptive Interleaved Visual Reasoners
by: Liu, Qingyang, et al.
Published: (2026)

Graph Integrated Multimodal Concept Bottleneck Model
by: Lin, Jiakai, et al.
Published: (2025)

CryoHype: Reconstructing a thousand cryo-EM structures with transformer-based hypernetworks
by: Gu, Jeffrey, et al.
Published: (2025)

Transductive Visual Programming: Evolving Tool Libraries from Experience for Spatial Reasoning
by: Wu, Shengguang, et al.
Published: (2025)

Connect, Collapse, Corrupt: Learning Cross-Modal Tasks with Uni-Modal Data
by: Zhang, Yuhui, et al.
Published: (2024)

VideoAgent: Long-form Video Understanding with Large Language Model as Agent
by: Wang, Xiaohan, et al.
Published: (2024)

Representation Forcing for Bottleneck-Free Unified Multimodal Models
by: Wang, Yuqing, et al.
Published: (2026)

Depth-guided NeRF Training via Earth Mover's Distance
by: Rau, Anita, et al.
Published: (2024)

Unleashing Perception-Time Scaling to Multimodal Reasoning Models
by: Li, Yifan, et al.
Published: (2025)

COOPER: A Unified Model for Cooperative Perception and Reasoning in Spatial Intelligence
by: Zhang, Zefeng, et al.
Published: (2025)

NegVQA: Can Vision Language Models Understand Negation?
by: Zhang, Yuhui, et al.
Published: (2025)

On the Perception Bottleneck of VLMs for Chart Understanding
by: Liu, Junteng, et al.
Published: (2025)

Perception, Reason, Think, and Plan: A Survey on Large Multimodal Reasoning Models
by: Li, Yunxin, et al.
Published: (2025)

Anny-Fit: All-Age Human Mesh Recovery
by: Bravo-Sánchez, Laura, et al.
Published: (2026)

V-GRPO: Online Reinforcement Learning for Denoising Generative Models Is Easier than You Think
by: Tang, Bingda, et al.
Published: (2026)

Video-STaR: Self-Training Enables Video Instruction Tuning with Any Supervision
by: Zohar, Orr, et al.
Published: (2024)

Template-Free Single-View 3D Human Digitalization with Diffusion-Guided LRM
by: Weng, Zhenzhen, et al.
Published: (2024)

Towards Faithful Multimodal Concept Bottleneck Models
by: Moreau, Pierre, et al.
Published: (2026)

Fine-tuning MLLMs Without Forgetting Is Easier Than You Think
by: Li, He, et al.
Published: (2026)

Apollo: An Exploration of Video Understanding in Large Multimodal Models
by: Zohar, Orr, et al.
Published: (2024)

Bridging Perception and Reasoning: Token Reweighting for RLVR in Multimodal LLMs
by: Lu, Jinda, et al.
Published: (2026)

Perception-Aware Multimodal Spatial Reasoning from Monocular Images
by: Cheng, Yanchun, et al.
Published: (2026)

GeoSense: Internalizing Geometric Necessity Perception for Multimodal Reasoning
by: Liu, Ruiheng, et al.
Published: (2026)