AI 情报文章归档

Three months ago, almost every AI coding tool had the same deal: pay a flat monthly fee, use it until...

This paper presents Murano, an open source framework for designing, running, and reproducing mechanistic interpretability studies of large language models, intended for researchers across disciplines. These studies often combine loading, r…

Large language model-based multi-agent systems are evolving from fixed interaction topologies toward dynamically orchestrated Agent Swarms. However, existing benchmarks are still largely based on single-agent or general-purpose agent tasks…

Production failed on Friday. My final transcript looked clean. The agent answered, cited sources, and...

I'd been checking my blog's search traffic by hand -- log into Search Console, click the property,...

Antarctic sea ice concentration (SIC) forecasting is an important yet challenging task due to the coexistence of complex spatial structure, long-range temporal dependencies, and strong seasonal variability. Conventional convolution-based m…

Multimodal Large Language Models (MLLMs) are increasingly deployed in multi-image scenarios requiring complex reasoning across visual contexts. However, current MLLMs remain fundamentally limited by object hallucination-generating plausibl…

Embedding methods have become popular due to their scalability on link prediction and/or triple classification tasks on Knowledge Graphs (KGs). Embedding models are trained relying on both positive and negative samples of triples. However,…

LLM agents need to sustain goal-consistent reasoning across long multi-turn interactions under strict resource constraints. However, as the multi-turn context accumulates, it can destabilize the underlying LLM's internal representation of …

Removing a specific individual's information from multimodal large language models (MLLMs) is often needed after deployment, but existing methods rely on a retain set, which is hardest to obtain at that point, and rebuilding it recreates t…

乌尔苏拉·冯德莱恩近日表示，她希望欧洲成为全球首个AI大陆。 韩国展示了如何抢先一步实现这一目标。 韩国正为每位公民提供免费、无限制的国产AI使用权限，并将其整合进医疗、住房、税务、教育和公共服务。 政府提供算力，本土企业提供模型和应用，公民获得普遍访问权。 这才是AI优先大陆的真正模样：AI作为公共基础设施。

ZHO 预测"AI 味"将成被追逐的风格，认为带 AI 感的美女脸比真人脸更吸引注意力。其机制是异常感触发"这是 AI 吗"的检测，劫持注意力优先于信息接收，使"AI 味"成为注意力战争中的特殊武器，并形成可消费的"诡异"，调用时代认知焦虑。

Language models can answer from precomputed memory, a model's saved reading of a body of material, reused across requests instead of read again at each. This paper maps where that practice preserves correctness and the conditions under whi…

Reliable uncertainty estimation is a crucial requirement for deploying large language models (LLMs) and vision-language models (VLMs) in safety-critical settings, especially when the model parameters are not accessible (black-box). We prop…

OpenShot 4.0 正式发布，带来录制视图、调色视图和本地 AI 遮罩等重大工作流升级。用户可直接录制麦克风、屏幕、摄像头和系统音频，并通过色轮、曲线、LUT 及专业视频示波器进行调色。新增 10 种特效，AI 遮罩模型完全本地运行，无需云服务或订阅。

While self-supervised approaches to reinforcement learning have achieved strong results by learning representations of states and actions, a key open question is the time scale over which actions should be modeled. Departing from the stand…

谷歌 AI 应用 Dreambeans 已取消付费订阅要求，并在美国开放免费测试。该应用是一款高度个性化的信息流应用，每天早晨提供数量有限的 AI 插图式新闻，而非无限浏览。Dreambeans 依托用户授权的 Gmail、日历、谷歌相册、YouTube、搜索记录等谷歌应用信息，生成地理位置推荐、生活随笔等个性化内容。

Optimizing absorption, distribution, metabolism, and excretion (ADME) is an important part of small molecule drug discovery. Many machine learning models have been built to predict ADME properties to facilitate this optimization process, b…

Timely post-disaster building damage assessment from satellite imagery is a critical engineering decision support task, yet it remains constrained by class imbalance, ambiguous intermediate damage states, and limited cross-event transferab…

Reinforcement learning (RL), particularly RL with Verifiable Rewards (RLVR), has recently emerged as a central paradigm for enhancing large language models' (LLMs) reasoning abilities, demonstrating remarkable effectiveness across reasonin…

DeepSeek 于 8 月 31 日在 Hugging Face 开源首个多模态模型 DeepSeek-V4-Flash-Vision-Exp，采用 MIT License，公开模型文件、Tokenizer、Prompt Encoding 参考实现及最小化 PyTorch 推理实现。

As language-model compute continues to scale, high-quality training data is becoming an increasingly important bottleneck. Conventional next-token prediction supervises what follows a context but leaves the intermediate reasoning behind th…

http://DIT.ai 已开放其 API 层的公共访问权限，允许用户在一个竞争性提供商市场中路由模型调用。 > 提供商按价格-质量评级排名。 > 它通过一个兼容 OpenAI 的密钥覆盖 50 多个文本、图像和视频模型。 > 设置仅需一个基础 URL 和密钥。 > 它包含提供商之间的自动故障转移开关。

自由智能必将胜利 （引用 @julien_c：开源AI必须获胜）

I wanted to find a free LLM API for small tools and experiments. There are many lists online, but...

华为在2026全球数据存储教授论坛揭晓第六届"奥林帕斯奖"，7支国内外科研团队分获1项奥林帕斯奖与6项先锋奖。最高奖由智能存储与计算重点实验室团队凭"层次化大内存的I/O优化与高效组织技术研究"方案斩获，该方案实现XPU下高性能数据访问并提升系统推理效率。本届赛事共吸引全球30所高校及机构的160多位学者申报。

Collecting natural-language referring expressions along with region annotations, such as masks or boxes, is a major bottleneck in visual grounding (VG), as annotators must write descriptions that distinguish target regions from visually si…

Synthetic data is increasingly used to train large language models (LLMs), yet its security implications remain poorly understood. Prior work on subliminal learning suggests that models can inherit behavioral traits from seemingly unrelate…

可灵以API形式接入百度秒哒无代码平台，累计调用超2496万次。开发者将可灵视频与图片生成能力直接嵌入应用，或批量生成视觉素材搭建世界观，如Movo调用可灵3.0Omni与3.0Turbo生成动态影像，互动文学游戏《奥德赛-归乡之歌》的角色与场景均由可灵生成。近期上线的Agent Market提供一键换装、商品成片等开箱即用Agent，同样开放API。

Organizing unstructured feedback text into hierarchical taxonomy is a fundamental challenge in NLP, particularly in domains where feedback arrives at massive scale in varied forms such as reviews, transcripts, and surveys. Existing approac…

Model Context Protocol servers make tools easier to discover and integrate, but discovery is only the...

Large language models are increasingly capable in general, but their utility can remain modest in niche or understudied areas. One approach to address this limitation is to specialise existing models through additional training on target-d…

With the development of large language models (LLMs), generative retrieval is becoming increasingly important in e-commerce scenarios. Current mainstream approaches typically use a two-stage training strategy: first train a product embeddi…

燧原科技公告科创板 IPO 发行价格定为 142.18 元 / 股，发行 4303.5173 万股，预计募集资金总额 61.19 亿元。该公司尚未盈利，对应 2025 年摊薄后静态市销率 61.80 倍，低于同行业可比公司平均水平。网上网下申购日为 2026 年 9 月 2 日，募资将用于第五代和第六代 AI 芯片研发及产业化等项目。

卡特彼勒正将采矿领域积累的自动化经验拓展至建筑工地等动态环境，推出 Cat AI Assistant 供技术人员通过语音调取维修流程和排查故障。该公司依托全球约 160 万台联网设备及超 16 PB 结构化数据，并利用 AI 智能体升级代码和生成数字孪生模型。未来五年，卡特彼勒计划投入 1 亿美元对 11.8 万名员工开展 AI、自动化和机器人技术培训。

Robust estimation is a core computer vision task frequently tackled using sample consensus. However, traditional methods suffer from inefficient sampling as they struggle to identify effective minimum sets before hypothesis evaluation. To …

I rewrote the system prompt fourteen times. Version 3 was "be thorough." Version 7 was a 900-word...

Valuation Guide There are only two ways anyone prices software. What it earns, or what it cost to...

智谱 2026 年上半年实现营业收入 9.54 亿元，同比增长 399.7%，云端部署收入占比提升至 86.5%，开放平台及 API 业务收入同比增长超 27 倍。期内归母净亏损 20.71 亿元，同比收窄 12.1%，经调整净亏损率同比收窄 3.5 倍。MaaS 平台用户数超 740 万，较年初增长 144%，付费日活用户增长 603%。

太棒了