NeurIPS 2021 accepted papers on Video
49 of the 2,508 papers accepted at NeurIPS 2021 (2.0%) are on Video, compared with 0.8% at NeurIPS 2020. They fall into 4 topics.
See NeurIPS 2021 in the interactive map
Submitting to NeurIPS? Checklist →
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Video segmentation and tracking · 26 papers
- Rethinking Space-Time Networks with Improved Memory Coverage for Efficient Video Object Segmentation (132 citations)
- Deep Contextual Video Compression (110 citations)
- TokenLearner: Adaptive Space-Time Tokenization for Videos (66 citations)
- Space-time Mixing Attention for Video Transformer (58 citations)
- Video Instance Segmentation using Inter-Frame Communication Transformers (58 citations)
- MAU: A Motion-Aware Unit for Video Prediction and Beyond (57 citations)
- CCVS: Context-aware Controllable Video Synthesis (30 citations)
- Exploring Cross-Video and Cross-Modality Signals for Weakly-Supervised Audio-Visual Video Parsing (28 citations)
- Drop-DTW: Aligning Common Signal Between Sequences While Dropping Outliers (26 citations)
- The Emergence of Objectness: Learning Zero-shot Segmentation from Videos (15 citations)
- Clockwork Variational Autoencoders (11 citations)
- SIMONe: View-Invariant, Temporally-Abstracted Object Representations via Unsupervised Video Decomposition (9 citations)
and 14 more in the interactive map.
Action recognition · 13 papers
- VATT: Transformers for Multimodal Self-Supervised Learning from Raw Video, Audio and Text (333 citations)
- Keeping Your Eye on the Ball: Trajectory Attention in Video Transformers (49 citations)
- Contrastive Learning of Global and Local Video Representations (20 citations)
- Low-Fidelity Video Encoder Optimization for Temporal Action Localization (16 citations)
- Long Short-Term Transformer for Online Action Detection (11 citations)
- Temporal-attentive Covariance Pooling Networks for Video Recognition (10 citations)
- Unsupervised Motion Representation Learning with Capsule Autoencoders (10 citations)
- MOMA: Multi-Object Multi-Actor Activity Parsing (9 citations)
- Shifted Chunk Transformer for Spatio-Temporal Representational Learning (7 citations)
- Relational Self-Attention: What's Missing in Attention for Video Understanding (6 citations)
- Spatiotemporal Joint Filter Decomposition in 3D Convolutional Neural Networks (4 citations)
- Action-guided 3D Human Motion Prediction (3 citations)
and 1 more in the interactive map.
Video-language understanding · 8 papers
- Detecting Moments and Highlights in Videos via Natural Language Queries (60 citations)
- MERLOT: Multimodal Neural Script Knowledge Models (54 citations)
- VALUE: A Multi-Task Benchmark for Video-and-Language Understanding Evaluation (41 citations)
- End-to-end Multi-modal Video Temporal Grounding (21 citations)
- Learning from Inside: Self-driven Siamese Sampling and Reasoning for Video Question Answering (18 citations)
- CLIP-It! Language-Guided Video Summarization (15 citations)
- A Bilingual, OpenWorld Video Text Dataset and End-to-end Video Text Spotter with Transformer (11 citations)
- Dynamic Visual Reasoning by Learning Differentiable Physics Models from Video and Language (10 citations)
Video generation · 2 papers
- FakeAVCeleb: A Novel Audio-Video Multimodal Deepfake Dataset (23 citations)
- Neural Dubber: Dubbing for Videos According to Scripts (2 citations)
