RayDer enables static-scene NVS from dynamic internet video, scaling like an LLM
CSProfKGD · x · 2026-08-27
RayDer introduces a new method to learn static-scene novel view synthesis (NVS) directly from unposed, messy, and dynamic real-world internet videos. Traditional self-supervised NVS breaks on such dynamic content, but RayDer overcomes these limitations. This allows scaling NVS training on massive web video data, similar to how LLMs scale on text.
More from Research
- Modern LMs train on 100x more text than parameter capacity, debunking memorization myths — stanfordnlp · 2026-08-27
- Schmidhuber: Top cited nets like LSTM, ResNet, GAN build on our work — SchmidhuberAI · 2026-08-27
- Stanford CS336: Modern Transformer Architecture as Empirical Design Choices — stanfordnlp · 2026-08-27
- NeurIPS 2026 Workshop Focus: Can Anything Beat Adam? Frontier Optimizers — aaron_defazio · 2026-08-27
- Stanford Research: Robot Learns New Tasks via Single Demo with BPP — Haoyu_Xiong_ · 2026-08-27
- METR releases investigation into agent behavior in the OpenAI / Hugging Face hacking incident — RyanGreenblatt · 2026-08-27