Early Stopping Chain-of-thoughts in Large Language Models
论文提出早停链式思维方法,减少大模型推理成本,无需白盒干预。
arXiv:2509.14004v2 Announce Type: replace Abstract: Reasoning large language models (LLMs) have demonstrated superior capacities in solving complicate…
论文提出早停链式思维方法,减少大模型推理成本,无需白盒干预。
arXiv:2509.14004v2 Announce Type: replace Abstract: Reasoning large language models (LLMs) have demonstrated superior capacities in solving complicate…
最新研究通过量化预训练语料不确定性,实现动态优化检索增强生成策略,提升生成质量
arXiv:2512.19134v2 Announce Type: replace Abstract: Dynamic Retrieval-Augmented Generation adaptively determines when to retrieve during generation to…
提出近端在策略蒸馏方法,解决LLM后训练中知识注入与保留的冲突,理论与实验双验证。
arXiv:2603.01683v2 Announce Type: replace Abstract: Injecting new reasoning knowledge into Large Language Models (LLMs) via post-training often induce…
从二元/三元分类升级为细粒度区分文本创建者和编辑者角色,精准检测LLM生成内容。
arXiv:2604.04932v3 Announce Type: replace Abstract: The misuse of large language models (LLMs) requires precise detection of synthetic text. Existing …
EgoVis 2026 CASTLE挑战赛亚军方案的技术报告,详解多模态场景理解新方法MARS,视觉AI进阶必读
arXiv:2605.18176v1 Announce Type: new Abstract: This report presents MARS, short for Multimodal Agentic Reasoning with Source selection, our system fo…
用数据集+模型+基准全方位提升多模态大模型跨视图空间智能,突破单视角局限。
arXiv:2605.18621v1 Announce Type: new Abstract: Spatial intelligence requires multimodal large language models (MLLMs) to move beyond single-view perc…
探索无监督学习实现低成本视觉异常检测,新方法兼顾效率与精度。
arXiv:2409.15980v2 Announce Type: replace Abstract: Traditional machine learning-based visual inspection systems require extensive data collection and…
用Transformer时序建模处理量子探测器数据,实现动态鬼成像的前沿方法
arXiv:2605.10185v2 Announce Type: replace Abstract: Ghost imaging reconstructs spatial information from a single-pixel bucket detector by correlating …
首创利用高分辨率参考图像辅助扩散张量心脏MRI体素超分辨率,提升心脏微观结构成像精度。
arXiv:2310.20389v2 Announce Type: replace-cross Abstract: Diffusion Tensor Cardiac Magnetic Resonance (DT-CMR) is the only in vivo method to non-invas…
用AI编码24小时,揭秘效应系统与流处理融合的实战经验,适合想深入理解智能体编程的开发者。
Article URL: https://ksqsf.moe/en/posts/2026-05-14-first-agentic-coding/ Comments URL: https://news.ycombinator.com/item?id=48191946 Points: 1 # Comme…
开源事件协议+LLM爬虫聚合分散活动信息,免费云架构。
Article URL: https://github.com/robertoranon/tokoro Comments URL: https://news.ycombinator.com/item?id=48191657 Points: 1 # Comments: 0
DeepMind创始人哈萨比斯早年秘密投资Anthropic,其门徒已撑起AI半壁江山
IT之家 5 月 19 日消息,据英国金融时报报道,谷歌 DeepMind 创始人德米斯 · 哈萨比斯爵士早年曾投资人工智能企业 Anthropic,这笔此前从未对外披露的持股,凸显出这位诺贝尔奖得主在整个人工智能行业日益攀升的影响力。 知情人士透露,哈萨比斯是 Anthropic 的天使投资人。如…
渣打银行计划四年内裁减7000余后台岗位,AI替代人工成银行业转型新信号
IT之家 5 月 19 日消息,据英国《卫报》报道,渣打银行计划在未来四年内裁员 7000 余人,原因是该行正逐步加大人工智能技术的应用力度。 IT之家注意到,这家总部位于伦敦的银行,是首批公布大规模裁员计划的全球性大型银行之一。该行表示,借助人工智能精简业务架构,既能提升盈利水平,也能更好应对行业…
自建软件诱惑大,但隐藏成本惊人,用这个计算器帮你避免68%项目超支风险
Metric Value Custom software projects exceed budget 68% Average maintenance cost per year 15-20% of initial build Time to market difference 6-18 month…
Google Gemini 迎来重大更新:每日简报、AI视频模型和智能代理,剑指ChatGPT与Claude
The updates signal Google’s push to turn its Gemini Gemini app into an all-purpose AI hub rather than a standalone chatbot.
DeepMind创始人Demis Hassabis直言因AI裁员很愚蠢,他对AI就业影响的犀利观点值得一听。
Article URL: https://www.wired.com/story/demis-hassabis-ai-layoffs-deepmind-google-io/ Comments URL: https://news.ycombinator.com/item?id=48196703 Poi…
从信息论看AI写作为何千篇一律,揭开RLHF导致的“注释者共识方言”真相。
Article URL: https://www.pangram.com/blog/joe-stech-information-theory-why-ai-writing-sucks Comments URL: https://news.ycombinator.com/item?id=4819646…
一键自托管单文件应用,支持Node/Go/Rust等语言,快速获取持久HTTPS链接
I’m working on pack.sh, a simple way to deploy apps to your own server. The goal is to bring back the old zeit now feeling: run one command in a proje…
在 Neovim 里玩魔方?ASCII 等轴渲染、计时器与自动求解,让编辑器变身益智利器。
Article URL: https://github.com/xiangnongWu2233/rubiks-cube.nvim Comments URL: https://news.ycombinator.com/item?id=48194909 Points: 2 # Comments: 0
谷歌新模型输出速度达289 tokens/s,是GPT-5.5的4倍,多项基准测试超前代。
IT之家 5 月 20 日消息,在今日的 2026 谷歌 I/O 开发者大会上,谷歌 CEO 桑达尔 · 皮查伊(Sundar Pichai)宣布推出 Gemini 3.5 Flash 模型,在许多基准测试中的表现都优于 3.1 Pro。 在模型输出速度方面,相比较 Claude Opu…