AI 情报文章归档

Stanford 论文提出经验性信息设计框架，在 10 个前沿 LLM 上对比不同披露粒度下的多智能体交易、议价与拍卖表现。

Agent memory can preserve mistakes. Anthropic's transcript-backed dreaming repairs notes, while contract-page ownership and exact quotations protect facts.

“Perfection is achieved, not when there is nothing more to add, but when there is nothing left to...

Why Andrew 'Red' Brogdon's insight on client-side functions is critical for Generative UI—and how bloc_signals_genui turns catalog functions into 0ms synchronous reactive signals.

Google 新论文发现，与会反驳的 AI 队友练习的学生写出的商业计划优于使用普通 AI 工具的学生。研究基于 Google Research 平台 Vantage，AI 队友由 Gemini 3.1 Pro 驱动，会提出有缺陷的想法、表达分歧并陈述错误说法，但不替学生完成任务。

Google is no longer describing AI Overviews, AI Mode and traditional Search as separate experiences....

Change impact analysis (CIA) helps developers identify code entities impacted by their initial changes. While several approaches for CIA were proposed, they still struggle with either low precision or low recall, thereby increasing manual …

Here I am sitting at home on a rainy Saturday, so I thought I'd share my story. It's nothing special,...

Options market makers absorb client order flow and must hedge the resulting inventory risk under trading frictions. We study whether an option hedging reinforcement learning (RL) agent trained on historical market data outperforms three al…

Cyber-physical smart-grid detectors are typically treated as black boxes and validated exclusively on clean, temporally shuffled datasets, which risk overly optimistic performance claims. This work presents a multi-testbed evaluation study…

This paper introduces COUPLED FLOW, a new neural operator learning framework for multi-physics PDE systems that captures the co-evolution of multiple physical processes by modeling their cross-process interactions. Unlike existing neural o…

Data transformation pipelines shape the training distribution in computer vision, but designing them is a computationally expensive problem. We investigate whether large language models can enable efficient discovery by synthesizing execut…

The field of mechanistic interpretability is well-suited to collaborative projects between mathematicians and AI researchers. This work describes one such project. We trained a small transformer to learn the language of the Coxeter group a…

Independent rank-tracking data identified a significant Google results anomaly in September and...

Researchers increasingly estimate political behavior based on race and gender from voter files, where race is imputed from surnames and geography. This only works where surnames suggest group membership, an assumption violated where surnam…

Existing evaluations of AI models on high-finance work focus on retrieving answers from financial documents and producing numerically correct outputs. Real private equity (PE) and investment banking (IB) work also depends on choosing betwe…

Binary decision-making under uncertainty arises across a wide range of applications, including energy systems scheduling, transportation, and workforce planning. While machine learning has been increasingly applied to accelerate or replace…

MarkTechPost 发布一篇教程，讲解如何用 llama.cpp 运行 NVIDIA Nemotron Labs 的 Dory 1.0 4B 循环推理模型。

Cerebras 发布了一段视频，展示 Kimi K3.1 在其晶圆级引擎上以每秒 2000+ tokens 的速度运行。 --- （完整视频见 Cerebras 的 YouTube 频道，链接在评论中。）

SemiAnalysis 团队 9 月实地探访内蒙古乌兰察布的 Galaxy 数据中心园区，规划 IT 容量约 800MW，与 Stargate Abilene 核心园区规模相当，首期 120MW 在一年内以约美国五分之一的成本交付。

In Bangla, a main verb is often followed by one of a small set of vector verbs. These verbs lose their own meaning and add a shade of meaning instead, such as completion. For example, likhe phello means that someone finished writing, not t…

Fixed prompt-based language models face challenges in open-domain conversations, where versatile topics require different expert advice or demonstrations to maximize their universal performance. Motivated by the dynamic prompt studies in l…

Detectors of AI-generated text are usually tuned and tested on passages of several hundred tokens, yet many real cases are short. We study score-based detection as a function of passage length. We give a closed form for how the false-posit…

Google can be prevented from fetching certain JavaScript resources during page rendering when those...

牛津 BOLD 团队论文研究 LLM 智能体的零样本协同（ZSC），在 9 个小型合作博弈中测试基于 GPT-5.2、Gemini 3 Pro 等模型的文本和代码智能体。

Real-world deployment of federated learning (FL) requires resilience against adversarial clients and data privacy. Byzantine fault captures the worst-case behavior of adversarial clients, while differential privacy (DP) provides statistica…

Standard ensemble methods can improve robustness by mitigating individual model stochasticity, but they frequently fail to generalize beyond the training distribution, leaving systems vulnerable to out-of-distribution (OOD) data. We introd…

Pre-training large parametric "Vision-Language-Action" (VLA) models on large demonstration datasets has recently yielded promising results for general-purpose manipulation. However, these models still struggle under even small domain shift…

Human-AI complementarity captures the idea that a human assisted by an AI system can outperform both the human and the AI acting alone. In prediction tasks, this improvement is assessed against ground truth labels produced through a refere…

Large Language Models (LLMs) exhibit approximate recall: they retain coarse associations about past events while losing fidelity on specific details. Past research on LLMs has quantified different models' forecasting ability in real-world …

Lighthouse 13.6.0 introduces changes to the simulation technology behind Lighthouse lab testing. That...

Much of the automation handed to language-model agents is procedural: each step needs little intelligence, but a long unattended run must follow a skill or SOP stated once at the start. We show that agents drift from such procedures over l…

Assessing climate risks requires projections of daily temperature distributions, including rare and spatially widespread extremes. Climate simulations are costly and biased, while correcting them toward observations risks suppressing inter…

Continual model merging incrementally integrates task-specific models without revisiting past models or data. However, repeated merging entangles past task components and distorts their singular geometry, making balanced task-wise allocati…

Big models are where inference engines show off. There's a lot of room to win. 27B is harder. The...

The New York Times has sued Brave Software, alleging that the browser and search company copied Times...

Submitted for the Flip the Switch Hackathon by WeMakeDevs & DevTools Academy (Powered by Switch...

据《金融时报》援引投资者的消息，Anthropic 计划在感恩节前 IPO，拟募资 800–1000 亿美元，估值 1.25–1.75 万亿美元，低于最乐观投资者期望的 2 万亿美元，但高于当前 9650 亿美元。

357 seconds. That's how long the median request waited for its first token when we ran GLM-5.3 on...

Andrej Karpathy 基于本周 Tristan 公布的 YouTube 频道分级清单，用 Claude Code 一次生成 TristanTube，收录 4.7 万个高技术含量视频并附带 TV 模式，全程仅用 3 分钟，站点在 karpathy.ai/tristan-tube/。此前 Tristan 也发布了用 Claude 从 20 年 HN 4300 万条评论中筛出 6454 个视频的 HackerTube，Karpathy 以此再次示范此类项目已可在极短时间…