1
Show HN: A extension for fast research and citations
浏览器扩展，划词即得APA/MLA等引用，八向径向菜单高效管理研究资料。
Article URL: https://github.com/fluryjanis/ResearchWheel Comments URL: https://news.ycombinator.com/item?id=49264396 Points: 1 # Comments: 0
2
Why KV Cache Matters — How MQA, GQA, and MLA Make LLM Inference Faster
深度拆解KV缓存与MQA、GQA、MLA原理，揭示大模型推理加速的关键技术。
LLMs generate text one token at a time. That sounds simple. But without KV Cache, every new token would repeat a lot of old work. That is why inferenc…
3
Homomorphic Encryptions for Privacy Preserving Vision
同态加密打破隐私壁垒，让敏感数据也能安全享受云端视觉AI服务。
arXiv:2606.25216v1 Announce Type: cross Abstract: Legal requirements might prevent organizations from sharing sensitive data like medical or financial…
4
Stochastic Penalty-Barrier Methods for Constrained Machine Learning
提出随机惩罚-障碍方法，解决约束机器学习优化难题，理论创新与算法实践兼备。
arXiv:2605.18618v1 Announce Type: new Abstract: Constrained machine learning enables fairness-aware training, physics-informed neural networks, and in…