From SRA to Self-Flow: Data Augmentation or Self-Supervision?
Research investigates the mechanisms behind self-alignment methods in diffusion transformers, finding that performance improvements stem primarily from data augmentation along the…
Hugging Face · Daily Papers
·Dengyang Jiang, Mengmeng Wang
·
·▲ 10 upvotes
Este artigo está em destaque na seleção diária de papers do Hugging Face, curada pela comunidade de pesquisa em IA.
Autores: Dengyang Jiang, Mengmeng Wang, Harry Yang, Jingdong Wang
- 10 upvotes da comunidade
- Temas: representation alignment, diffusion transformer, self-alignment, SRA, Self-Flow, dual-time scheduling
Resumo
Resumo original (em inglês), extraído do paper:
Research investigates the mechanisms behind self-alignment methods in diffusion transformers, finding that performance improvements stem primarily from data augmentation along the noise dimension rather than token interactions between noise levels.Onde ler
// relacionados
Leia também
Blog
O complicado problema do Claude Code com a China envolve proibições dos dois lados do Pacífico
Blog
AI Security Institute do Reino Unido descobre que benchmarks padrão subestimam sistematicamente o que agentes de IA realmente conseguem fazer
Dataset
ByteDance-Seed/EdgeBench
Blog