#optimization
3 articles
逆向擴散的一階駐點性：連結優化與生成採樣的數學機制
This research establishes a first-order optimization theory for diffusion models, proving that SDE-based reverse Langevin diffusions contract Fisher divergences exponentially under strongly convex noising.
直接回饋對齊（DFA）中的共模崩潰與恢復機制
This study uncovers the 'common-mode collapse' in Direct Feedback Alignment (DFA) where shared error components cause neuron saturation, and proposes effective mitigation strategies.
程式碼代理人的精簡文件有用嗎？研究揭示驚人否定結果：有了源碼，文件反而成多餘
This study investigates if compact natural-language documentation helps coding agents resolve software issues, finding that despite building high-fidelity documentation, it fails to improve real-world issue resolution over using raw source code and issue descriptions alone.