Trimming the Long-Tail of Visual World Modeling Evaluation
Current visual world models demonstrate limited generalization beyond common physical interactions, struggling with rare and irregular scenarios despite achieving realism on standa…
Hugging Face · Daily Papers
·Bingxuan Li, Yining Hong
·
·▲ 35 upvotes
Este artigo está em destaque na seleção diária de papers do Hugging Face, curada pela comunidade de pesquisa em IA.
Autores: Bingxuan Li, Yining Hong, Cheng Qian, Hyeonjeong Ha, Jiateng Liu, Zhenhailong Wang
- 35 upvotes da comunidade
- Temas: visual world models, physical interactions, long-tailed distribution, image generation, video generation, world model evaluation
Resumo
Resumo original (em inglês), extraído do paper:
Current visual world models demonstrate limited generalization beyond common physical interactions, struggling with rare and irregular scenarios despite achieving realism on standard benchmarks.Onde ler
// relacionados
Leia também
Blog
Google launches Nano Banana 2 Lite for fast AI images and Gemini Omni Flash for video via API
Blog
Constrained Tabular Diffusion for Finance
Blog
DiffRGD: An Inference-Time Diffusion Guidance Through Riemannian Gradient Descent
Blog