ICLR 2025 accepted papers on Video
149 of the 3,703 papers accepted at ICLR 2025 (4.0%) are on Video, compared with 2.0% at ICLR 2024. They fall into 4 topics.
See ICLR 2025 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Video generation · 82 papers
- CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer (18 citations)
- CameraCtrl: Enabling Camera Control for Video Diffusion Models (16 citations)
- Diffusion Models Are Real-Time Game Engines (5 citations)
- Bridging the Data Provenance Gap Across Text, Speech, and Video (4 citations)
- Real-Time Video Generation with Pyramid Attention Broadcast (4 citations)
- Ctrl-Adapter: An Efficient and Versatile Framework for Adapting Diverse Controls to Any Diffusion Model (3 citations)
- Hallo2: Long-Duration and High-Resolution Audio-Driven Portrait Image Animation (3 citations)
- 4K4DGen: Panoramic 4D Generation at 4K Resolution (2 citations)
- Fantastic Copyrighted Beasts and How (Not) to Generate Them (2 citations)
- GameGen-X: Interactive Open-world Game Video Generation (2 citations)
- OpenVid-1M: A Large-Scale High-Quality Dataset for Text-to-video Generation (2 citations)
- Pyramidal Flow Matching for Efficient Video Generative Modeling (2 citations)
and 70 more in the interactive map.
Video-language understanding · 48 papers
- World Model on Million-Length Video And Language With Blockwise RingAttention (12 citations)
- One Hundred Neural Networks and Brains Watching Videos: Lessons from Alignment (4 citations)
- Shot2Story: A New Benchmark for Comprehensive Understanding of Multi-shot Videos (3 citations)
- Dense Video Object Captioning from Disjoint Supervision (2 citations)
- GUI-World: A Video Benchmark and Dataset for Multimodal GUI-oriented Understanding (2 citations)
- LongVILA: Scaling Long-Context Visual Language Models for Long Videos (1 citation)
- MMWorld: Towards Multi-discipline Multi-faceted World Model Evaluation in Videos (1 citation)
- Oryx MLLM: On-Demand Spatial-Temporal Understanding at Arbitrary Resolution (1 citation)
- Streaming Video Understanding and Multi-round Interaction with Memory-enhanced Knowledge (1 citation)
- TimeSuite: Improving MLLMs for Long Video Understanding via Grounded Tuning (1 citation)
- TOMATO: Assessing Visual Temporal Reasoning Capabilities in Multimodal Foundation Models (1 citation)
- Aligned Better, Listen Better for Audio-Visual Large Language Models
and 36 more in the interactive map.
Video segmentation and tracking · 13 papers
- SAM 2: Segment Anything in Images and Videos (264 citations)
- VEDIT: Latent Prediction Architecture For Procedural Video Representation Learning (1 citation)
- BaB-ND: Long-Horizon Motion Planning with Branch-and-Bound and Neural Dynamics
- Depth Any Video with Scalable Synthetic Data
- Dreamweaver: Learning Compositional World Models from Pixels
- Efficient Masked AutoEncoder for Video Object Counting and A Large-Scale Benchmark
- EgoExo-Gen: Ego-centric Video Prediction by Watching Exo-centric Videos
- Learning Spatial-Semantic Features for Robust Video Object Segmentation
- SMITE: Segment Me In TimE
- Animate Your Thoughts: Reconstruction of Dynamic Natural Vision from Human Brain Activity
- Local Patterns Generalize Better for Novel Anomalies
- PooDLe🐩: Pooled and dense self-supervised learning from naturalistic videos
and 1 more in the interactive map.
Action recognition · 6 papers
- ALBAR: Adversarial Learning approach to mitigate Biases in Action Recognition
- Grounding Video Models to Actions through Goal Conditioned Exploration
- Masked Temporal Interpolation Diffusion for Procedure Planning in Instructional Videos
- Video Action Differencing
- Action Sequence Augmentation for Action Anticipation
- Learning Video-Conditioned Policy on Unlabelled Data with Joint Embedding Predictive Transformer