微信内可能无法直接打开本站。请点右上角 ··· → 在浏览器打开,或复制链接后用系统浏览器访问。
XMT
信流 · 上滑连读 · 来源可核
IJCAI 2026 复盘局:中国人投了最多的稿,却到不了最多的场
CCF降级没能让中国学者少投,但一纸签证却让他们没法到场。作者丨幸丽娟 编辑丨岑 峰 德国不莱梅当地时间8月22日下午,IJCAI-ECAI 2026 迎来闭幕式。程序委员会主席 Diego Calvanese走上讲台,用一组详实的数据为这场为期一周的学术盛会画上句号。…
- 而站在这个节点上,IJCAI所呈现的不仅是论文与参会者的数字图谱,也是一张值得 AI学术圈细读的底稿。
- 01注册与投稿数量:中国两个“第一”的温差Calvanese 先公布了注册数据的地区分布:中国669人,占26%,排第一;德国667人,26…
- Calvanese 还特别解释了一句,德国注册量高,本地组委会的推广起了作用。
RSS 官方收录 · 可信分层展示
李飞飞押注的空间智能:生成的世界,如何「经得起折腾」?
AI 生成的大多数世界“中看不中用”,根本原因是 AI 少了一本“底账”:一份记住世界里有什么、每次行动改变了什么、并且能被检查和回滚的状态记录。作者丨刘紫东 编辑丨幸丽娟 李飞飞押注的空间智能,被视为世界模型的主流路线之一。…
- 这条路线的基本构想是,让 AI 从识别二维画面,走向理解、推理和生成可进入、可操作、可持续编辑的三维世界。
- 但真正的考验在于:一个“站得住”的世界,是怎样构建出来的?
- 要回答这个问题,不妨先回到一个更基础的追问:当 AI 的交付形态从“画面”变成“三维世界”,互动方式从“观看”变成“进入和操作”,到底意味着…
RSS 官方收录 · 可信分层展示
通用模型竟然比医疗专用模型更懂医疗?一篇 ACL 论文的两个反直觉发现 | GAIR Paper 123
Doubao 、DeepSeek 在三语数据集上的表现,普遍超过 GPT-4o 和 Gemini。作者丨幸丽娟 编辑丨岑 峰 医疗,是大模型落地最特殊的场景之一。特殊在哪里?特殊在一个很微小的错误,就可能导致致命的后果;也特殊在你不能用直觉去判断它"行"还是"不行"——你觉得医疗专用模型肯定比通用模型更懂医疗?…
- 你觉得英语训练出来的模型处理英语医疗数据肯定最好?
- 这两重"反直觉",正是纽约大学阿布扎比分校和阿布扎比克利夫兰诊所的研究团队在ACL 2026 Findings上发表的论文揭示的核心真相。
- 论文链接:https://aclanthology.
RSS 官方收录 · 可信分层展示
Hate Speech Classification In Roman Urdu: A Comparative Study On Parameter Efficient Fine-Tuning And Prompt Engineering
arXiv:2608.…
- 21408v1 Announce Type: new Abstract: Due to the widespread accessibili…
- Ro-man Urdu, a low-resource language used in Pakistan and among Urdu-s…
- This research aims to identify the most effective techniques for hate …
RSS 官方收录 · 可信分层展示
Spyre-Accelerated Retrieval-Augmented Generation on IBM LinuxONE: A Cloud-Native Architecture for Secure, High-Throughput Enterprise AI Inference
arXiv:2608.…
- 21393v1 Announce Type: new Abstract: Running large language models ins…
- IBM's Spyre accelerator PCIe inference card built for LinuxONE and the…
- In this paper we lay out a six-subsystem RAG architecture that runs en…
RSS 官方收录 · 可信分层展示
There Is No Neutral Harness: Modern LLM Leaderboards Are Manufactured by Config-Fragile Items
arXiv:2608.…
- 21382v1 Announce Type: new Abstract: Multiple-choice benchmarks fix th…
- Work on this harness sensitivity reports it as aggregate score varianc…
- We treat the evaluation harness of large language models (LLMs) as an …
RSS 官方收录 · 可信分层展示
RIACT: A Responsible AI System for Personalized Study Habit Tracking and Early Burnout Signal Detection in University Students
arXiv:2608.…
- 21379v1 Announce Type: new Abstract: Student burnout is highly prevale…
- A contributing factor is that students have little structured visibili…
- This paper presents RIACT (Record, Insight, Analyze, Coach, Track), a …
RSS 官方收录 · 可信分层展示
SchemaRouter: Field-Aware Tool Routing for Efficient Heterogeneous Agentic RAG
arXiv:2608.…
- 21375v1 Announce Type: new Abstract: Heterogeneous agentic retrieval-a…
- Exposing all tool descriptions to an LLM agent, or selecting tools onl…
- We present SchemaRouter, a lightweight routing layer that represents t…
RSS 官方收录 · 可信分层展示
NEEDLES and Brooks Brothers Merge Heritage Tailoring With Modern Streetwear
SummaryNEEDLES and Brooks Brothers have officially announced a collaborative capsule collection merging American menswear with Japanese designThe offering features tailored sport coats matching trousers check button-downs v-neck and mock neck sweaters silk ties and leather loafersThe collection drops via a two-stage release beginning September 2 followed by a wider rollout on September 9NEEDLES and Brooks Brothers have officially unveiled a collaborative capsule collection merging classic American tailoring with Japanese design perspectives.…
- Created in partnership with NEPENTHES founder Keizo Shimizu, the upcom…
- The extensive collection features tailored outerwear, knitwear, access…
- Highlight pieces include Sport Coats (¥107,800 JPY) and matching Trous…
RSS 官方收录 · 可信分层展示