Research map / Vision-language and multimodal / Cross-modal learning
Everything is a Video: Unifying Modalities through Next-Frame Prediction
G. Thomas Hudson, Dean Slack, Thomas Winterbottom, Jamie Sterling, Chenghao Xiao, Junjie Shentu, Noura Al Moubayed
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| limit evaluation to inputs of 20 tokens or fewer instead of truncating longer inputs | next-frame model · SST-2 | F1: 76.8 -> 80.0 | Improved |
| reformulate classification as next-frame prediction without pretraining vs ViT-base | next-frame model · CIFAR-10 | accuracy: 71.3 -> 89.1 | Improved |
| reformulate as next-frame prediction vs PCANet | next-frame model · CIFAR-10 | accuracy: 77.1 -> 89.1 | Improved |
| next-frame prediction vs pretrained ViT-huge | next-frame model · CIFAR-10 | accuracy: 99.5 -> 89.1 | Hurt |
| next-frame prediction vs pretrained SOTA | next-frame model · SST-2 | F1: 91.3 -> 76.8 | Hurt |
| next-frame prediction vs ResNet50 | next-frame model · TinyVIRAT | macro F1: 29.1 -> 30.4 | Improved |
| next-frame prediction vs AlexNet | next-frame model · AudioMNIST | accuracy: 95.8 -> 97.1 | Improved |
| next-frame prediction vs LSTM | next-frame model · CLEVRER (descriptive) | accuracy: 34.7 -> 52.5 | Improved |
| next-frame prediction vs prior tracker | next-frame model · LaSOT | IoU: 0.60 -> 0.63 | Improved |
| next-frame colorization vs ground-truth dataset colors | next-frame model · TinyVIRAT colorization | colorfulness: 70.6 -> 73.1 | Mixed |
Closest papers by meaning
- Emu: Generative Pretraining in Multimodality (ICLR 2024 · similarity 0.90)
- Mirasol3B: A Multimodal Autoregressive Model for Time-Aligned and Contextual Modalities (CVPR 2024 · similarity 0.90)
- Dissecting Multimodality in VideoQA Transformer Models by Impairing Modality Fusion (ICML 2024 · similarity 0.89)
- CREMA: Generalizable and Efficient Video-Language Reasoning via Multimodal Modular Fusion (ICLR 2025 · similarity 0.89)
- V^2Dial: Unification of Video and Visual Dialog via Multimodal Experts (CVPR 2025 · similarity 0.89)
- MSP: Modality Self-Play for Training Multimodal LLMs Without Paired Cross-Modal Data (NeurIPS 2026 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
