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
短闻
信流 · 上滑连读 · 来源可核
1 / 24
Human-AI Co-Interpretation for Responsible AI: A Hermeneutic Perspective
arXiv:2609.…
- 00334v1 Announce Type: new Abstract: Across law, education, policy ana…
- Yet a recurrent failure mode -- what I call \textit{interpretive mispl…
- In such settings, the risk is not only factual error but lost accounta…
RSS 官方收录 · 可信分层展示
The Irreversibility Budget: Fleet-Level Risk Accounting and Admission Control for Agent Operating Systems
arXiv:2609.…
- 00275v1 Announce Type: new Abstract: Fleets of LLM agents now external…
- Current controls check one effect at a time, so a fleet of individuall…
- We propose the irreversibility budget, a cumulative account of residua…
RSS 官方收录 · 可信分层展示
Autoresearch for Marketplace Catalogs: From Legacy Forms to AI-Native Matching
arXiv:2609.…
- 00274v1 Announce Type: new Abstract: Two-sided service marketplaces ar…
- Relying on inferred intent rather than fixed-form fields forces these …
- We present an autoresearch loop that generates this taxonomy, one occu…
RSS 官方收录 · 可信分层展示
The Answer Is Not the Argument
arXiv:2609.00264v1 Announce Type: new Abstract: Chain-of-thought monitoring is proposed for AI oversight, yet evaluations often provide monitors with a trusted reference answer.…
- We ask whether answer access improves reasoning verification or mainly…
- We collected 237 step-numbered solutions to 79 Humanity's Last Exam ph…
- The reference standard combined physicist annotations, an independent …
RSS 官方收录 · 可信分层展示
IFA 2026：添可诠释智能清洁的未来图景
“添可洗地机，定义每一代“不只是一句口号，更代表着添可不断突破创新、让生活更简单的长期承诺。作为智能洗地机品类开创者，添可将再次亮相2026年9月4日至8日举办的德国柏林国际消费电子展（IFA）——这是欧洲规模最大的消费电子与家电行业盛会。…
- 届时，添可将带来覆盖多个家居清洁品类的全新产品，进一步展现其推动智能清洁标准不断升级的决心。
- 自净即是未来：添可眼中的明日清洁愿景本届IFA上，添可带来了全新传播主题“The Future Cleans Itself”（自净即是未来）。
- 在这一主题下，添可描绘了一个更智能、更自动、也更简单的清洁未来。
RSS 官方收录 · 可信分层展示
Authority Bias in Conversational Search Engines for Academic Paper Recommendation
arXiv:2609.…
- 00248v1 Announce Type: new Abstract: Large Language Models (LLMs) are …
- We investigate authority bias: systematic preference for papers based …
- Holding title and abstract constant, we vary authority metadata across…
RSS 官方收录 · 可信分层展示
Hypotheses-Guided Self Distillation for Continual Personalization
arXiv:2609.…
- 00251v1 Announce Type: new Abstract: As people increasingly interact w…
- However, user preferences are rarely stated in full, and instead emerg…
- We introduce HypReflect, a reliable, scalable framework for continual …
RSS 官方收录 · 可信分层展示
Invalidation Contracts for Cross-Episode Agent Memory
arXiv:2609.…
- 00243v1 Announce Type: new Abstract: LLM agents that cache recovery su…
- Server-side data drift turns those cached fixes into silent failures, …
- We introduce invalidation contracts, a protocol layer that attaches ve…
RSS 官方收录 · 可信分层展示
Learning What to Retain: Gated-Memory Routing for Efficient Collaboration in Multi-Agent LLM Systems
arXiv:2609.…
- 00237v1 Announce Type: new Abstract: Large language model (LLM)-based …
- A central challenge is to adapt orchestration to the evolving collabor…
- Routing from the query alone cannot adapt to intermediate progress or …
RSS 官方收录 · 可信分层展示
ConvDeck: Conversational Paper-to-Slide Generation via Stage-Specific User Feedback
arXiv:2609.…
- 00226v1 Announce Type: new Abstract: Automatic academic paper-to-slide…
- Recent multi-agent systems partially acknowledge this through internal…
- However, these refinement processes either remain largely closed to th…
RSS 官方收录 · 可信分层展示
建立AI时代“度量衡”：OPPO联合OpenKG推出首个端侧AI记忆评测基准MobileMem
随着终端AI从“回答问题”走向长期理解用户和主动任务执行，记忆能力正成为决定AI个性化与主动服务水平的关键基础能力。近日，OPPO联合中文开放知识图谱（OpenKG）社区，推出行业首个面向端侧AI记忆能力的评测基准MobileMem。…
- 同时携手浙江大学、同济大学、东南大学等高校专家团队共同建设。
- 通过构建统一、开放的评测“标尺”， MobileMem让不同技术路线在同一标准下实现可比较、可迭代，加速端侧AI记忆技术发展。
- 这也意味着OPPO端侧AI记忆的持续投入，正从自身产品能力建设，进一步延伸至行业公共能力建设，积极推动技术开放合作与生态发展。
RSS 官方收录 · 可信分层展示
AI Should Not Only Be Helpful. It Should Be Contingent. Artificial Intimacy, Sycophancy, and the Future of Social Learning
arXiv:2609.…
- 00211v1 Announce Type: new Abstract: Conversational artificial intelli…
- This perspective introduces contingency, i.
- , the degree to which system responses vary with user behavior and its…
RSS 官方收录 · 可信分层展示
ReDeck: Step-Level Render-Grounded Refinement for Document-to-Slide Generation
arXiv:2609.…
- 00194v1 Announce Type: new Abstract: Document-to-slide generation is c…
- Recent slide agents adopt iterative reflection, but typically follow a…
- This delayed feedback makes local failures such as overflow, overlap, …
RSS 官方收录 · 可信分层展示
LLM-Driven Autonomous Vehicles Inherit Human Driver Biases in Pedestrian Yielding: Results and Implications From A New Benchmark
arXiv:2609.…
- 00192v1 Announce Type: new Abstract: Public trust in Autonomous Vehicl…
- While a recent trend in AV research involves using general purpose "co…
- Given that psychology studies have shown human driver biases exist, su…
RSS 官方收录 · 可信分层展示
Asymmetries in Spontaneous and Instructed Deception
arXiv:2609.00180v1 Announce Type: new Abstract: Large language models sometimes deceive users without being instructed to.…
- However, much of the study on deception in models involves instructed …
- We investigated the relationship between instructed and spontaneous (u…
- 1-70B-Instruct.
RSS 官方收录 · 可信分层展示
IMPACT: Attention Is the Interaction Map for Scalable Interaction-Aware World Model Training
arXiv:2609.…
- 00161v1 Announce Type: new Abstract: World models have made remarkable…
- Existing approaches address this limitation by constraining the genera…
- Obtaining these spatiotemporally dense representations typically requi…
RSS 官方收录 · 可信分层展示