Image de couverture de livre

On Discriminative vs. Generative classifiers: Rethinking MLLMs for Action Understanding

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Pang, Zhanzhong, Chatterjee, Dibyadip, Sener, Fadime, Yao, Angela
Format:	Preprint
Publié:	2026
Sujets:	Computer Vision and Pattern Recognition
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Decouple and Cache: KV Cache Construction for Streaming Video Understanding
par: Pang, Zhanzhong, et autres
Publié: (2026)

Don't Pause! Every prediction matters in a streaming video
par: Chatterjee, Dibyadip, et autres
Publié: (2026)

Context-Enhanced Memory-Refined Transformer for Online Action Detection
par: Pang, Zhanzhong, et autres
Publié: (2025)

Cost-Sensitive Learning for Long-Tailed Temporal Action Segmentation
par: Pang, Zhanzhong, et autres
Publié: (2025)

Long-Tail Temporal Action Segmentation with Group-wise Temporal Logit Adjustment
par: Pang, Zhanzhong, et autres
Publié: (2024)

On the Utility of 3D Hand Poses for Action Recognition
par: Shamil, Md Salman, et autres
Publié: (2024)

Memory-efficient Streaming VideoLLMs for Real-time Procedural Video Understanding
par: Chatterjee, Dibyadip, et autres
Publié: (2025)

X-MIC: Cross-Modal Instance Conditioning for Egocentric Action Generalization
par: Kukleva, Anna, et autres
Publié: (2024)

Rethinking Domain Generalization: Discriminability and Generalizability
par: Long, Shaocong, et autres
Publié: (2023)

SneakPeek: Future-Guided Instructional Streaming Video Generation
par: Hong, Cheeun, et autres
Publié: (2025)

GOBench: Benchmarking Geometric Optics Generation and Understanding of MLLMs
par: Zhu, Xiaorong, et autres
Publié: (2025)

PALM: A Dataset and Baseline for Learning Multi-subject Hand Prior
par: Fan, Zicong, et autres
Publié: (2025)

OST-Bench: Evaluating the Capabilities of MLLMs in Online Spatio-temporal Scene Understanding
par: Lin, Jingli, et autres
Publié: (2025)

Memory Helps, but Confabulation Misleads: Understanding Streaming Events in Videos with MLLMs
par: Zhang, Gengyuan, et autres
Publié: (2025)

Two Causes, Not One: Rethinking Omission and Fabrication Hallucinations in MLLMs
par: Si, Guangzong, et autres
Publié: (2025)

Condensing Action Segmentation Datasets via Generative Network Inversion
par: Ding, Guodong, et autres
Publié: (2025)

EgoExoBench: A Benchmark for First- and Third-person View Video Understanding in MLLMs
par: He, Yuping, et autres
Publié: (2025)

Spatial Preference Rewarding for MLLMs Spatial Understanding
par: Qiu, Han, et autres
Publié: (2025)

Aligning Generative Denoising with Discriminative Objectives Unleashes Diffusion for Visual Perception
par: Pang, Ziqi, et autres
Publié: (2025)

Fact :Teaching MLLMs with Faithful, Concise and Transferable Rationales
par: Gao, Minghe, et autres
Publié: (2024)

Universal Skeleton Understanding via Differentiable Rendering and MLLMs
par: Wang, Ziyi, et autres
Publié: (2026)

3D Spatial Understanding in MLLMs: Disambiguation and Evaluation
par: Chang, Chun-Peng, et autres
Publié: (2024)

F1: A Vision-Language-Action Model Bridging Understanding and Generation to Actions
par: Lv, Qi, et autres
Publié: (2025)

GAIA: Rethinking Action Quality Assessment for AI-Generated Videos
par: Chen, Zijian, et autres
Publié: (2024)

X-Stream: Exploring MLLMs as Multiplexers for Multi-Stream Understanding
par: Sun, Peiwen, et autres
Publié: (2026)

ODI-Bench: Can MLLMs Understand Immersive Omnidirectional Environments?
par: Yang, Liu, et autres
Publié: (2025)

Math Blind: Failures in Diagram Understanding Undermine Reasoning in MLLMs
par: Sun, Yanpeng, et autres
Publié: (2025)

Pose-RFT: Enhancing MLLMs for 3D Pose Generation via Hybrid Action Reinforcement Fine-Tuning
par: Li, Bao, et autres
Publié: (2025)

Coherent Temporal Synthesis for Incremental Action Segmentation
par: Ding, Guodong, et autres
Publié: (2024)

Context Tokens are Anchors: Understanding the Repetition Curse in dMLLMs from an Information Flow Perspective
par: Zhao, Qiyan, et autres
Publié: (2026)

OnlineTAS: An Online Baseline for Temporal Action Segmentation
par: Zhong, Qing, et autres
Publié: (2024)

Bridging Degradation Discrimination and Generation for Universal Image Restoration
par: Hu, JiaKui, et autres
Publié: (2026)

Rethinking Prior Information Generation with CLIP for Few-Shot Segmentation
par: Wang, Jin, et autres
Publié: (2024)

E-VAds: An E-commerce Short Videos Understanding Benchmark for MLLMs
par: Liu, Xianjie, et autres
Publié: (2026)

Open Eyes, Then Reason: Fine-grained Visual Mathematical Understanding in MLLMs
par: Zhang, Shan, et autres
Publié: (2025)

STI-Bench: Are MLLMs Ready for Precise Spatial-Temporal World Understanding?
par: Li, Yun, et autres
Publié: (2025)

AdaptVision: Dynamic Input Scaling in MLLMs for Versatile Scene Understanding
par: Wang, Yonghui, et autres
Publié: (2024)

Knowledge is Power: Advancing Few-shot Action Recognition with Multimodal Semantics from MLLMs
par: Xing, Jiazheng, et autres
Publié: (2026)

Rethinking generalization of classifiers in separable classes scenarios and over-parameterized regimes
par: Martinetz, Julius, et autres
Publié: (2024)

From Text to Pixel: Advancing Long-Context Understanding in MLLMs
par: Lu, Yujie, et autres
Publié: (2024)