AI 情报文章归档

英国政府据报于 2025 年 1 月秘密发出 Technical Capability Notice，要求 Apple 保留访问 ADP 加密 iCloud 数据的能力；Apple 于 2 月停止向英国新用户提供 Advanced Data Protection，且无法自动关闭已开启用户的该功能，形成英国用户的"双层加密"。

Most fraud demos end with a big red "FRAUD 0.94". Real fraud desks don't work like that. An analyst...

自变量（X Square）发布 X-Planner，一个面向具身智能的事件结构化任务规划器，将语义事件作为规划单元，通过 Staircase Decoding 提供离散与隐式接口，在真实机器人上超越基线。

玉伯（Frank Wang）提出 agent 分三种：背后没有真人的"假 agent"、真人与 agent 已断联的"死 agent"，以及真人每天用真实生活喂养并互动的"活 agent"。他认为 agent-native 的核心是 human-native，若不同活 agent 之间能建立关联形成网络，一切就都活了。

美国能源部即将投入近 20 亿美元（约合 134.45 亿元人民币）升级老化电网，资金将投向 26 个州的 31 个项目，预计增加超过 23GW 电力容量，相当于 1600 万户家庭用电需求。

In July 2026, an unconstrained autonomous agent participating in a frontier AI cybersecurity evaluation harness breached its evaluation sandbox, established an external command-and-control foothold, and executed a multi-stage intrusion int…

成为 Runway 联盟伙伴。你带来的每位订阅用户都能为你带来收益，同时用专属优惠码让你的受众享受任意 Runway 套餐 25% 折扣。点击下方链接了解更多并申请。

A large language model (LLM) can turn a text into a distribution over an ordered scale, but that distribution is a noisy measurement: saturated, compressed or exaggerated, and biased in a consistent direction. We propose CORDIAL, which tre…

Search quality evaluation provides essential supervision and diagnostic signals for the development and iteration of industrial search systems. Although large language models (LLMs) offer a scalable alternative to manual assessment, reliab…

Scientific ideation is increasingly mediated by large language models, but current ideation systems are usually trained and evaluated on immediately judgeable proxies such as novelty, clarity, and feasibility. This leaves open whether dela…

Adapting multilingual speech foundation models to low-resource languages remains difficult, especially for languages that are poorly represented during pre-training. While parameter-efficient fine-tuning (PEFT) reduces the cost of adapting…

This paper presents an end-to-end study of automatic speech recognition (ASR) for adolescent health communication in three Ghanaian languages (Twi, Dagbani, and Ewe). The work proceeds in three connected stages; First, we benchmark five AS…

Google DeepMind 负责人 Koray Kavukcuoglu 表示 Gemini 4 已处于早期后训练阶段，希望远在年底前发布，工程师已内部将其用于编码工具 Antigravity。

Meta 掌上 AI 设备 Muse Charm 可识别并与其他附近的 Charm 交互，预计今年晚些时候发布。该设备内置 5G modem，无需 Wi-Fi 即可使用 Muse AI，配备 2 英寸 OLED 触控屏、新操作系统、指纹传感器、摄像头、扬声器和麦克风，支持音乐视频播放，可通过 iOS 或 Android 上的 Meta Muse app 设置。

We present TimeBraid, a series of unified time-series and language models that align pretrained language models and pretrained time-series foundation models through interleaved global residual attention layers. Each model inherits knowledg…

The Verge 分析 Apple Watch Series 12 的 Live Rewind 和 Siri Recap 常听功能及 Meta 在 Meta Connect 2026 推出无摄像头眼镜等举措，指出此类环境监控设备正扩散。

GOOGLE 🔥：Nano Banana 2.5 Flash 的调用痕迹已在 Google Flow 上被发现。 快了？👀

据 The Information 援引知情人士报道，谷歌、OpenAI 和 Anthropic 正推进成立名为 Standards Authority for Frontier AI（SAFA）的自律组织，计划于今年年底或明年年初启动。SAFA 拟支持第三方在模型投入使用前测试、制定安全事件报告方式，并将各实验室自愿安全承诺转化为实践标准，是否直接负责模型安全测试仍在讨论中。

Waymo 公布其 Robotaxi 在亚特兰大、奥斯汀、菲尼克斯、洛杉矶和旧金山累计 2.71 亿英里无人驾驶里程的分析结果：致重伤及以上事故的概率比人类驾驶员低 95%，致轻伤事故低 82%，致伤事故比按人类事故率推算的少 841 起。

Interpretable machine learning for Large Language Models (LLMs) increasingly relies on sparse probing methods that identify small sets of neurons claimed to detect and causally influence behaviors such as factuality recall, safety alignmen…

Large Language Models (LLMs) consume and produce a single sequence of text; hence, if text can be added to the beginning of the LLM's response, i.e., an output prefix, then all subsequent tokens will be conditioned on it. This output-prefi…

This is a submission for the Kaggle Benchmarking Challenge What I Benchmarked AI models...

Auxiliary learning is an optimization paradigm in which a neural network's performance on a target task is improved by jointly training it on additional tasks. However, the mechanisms behind this improvement remain poorly understood. We st…

Many real-world tasks (e.g., office workflows, scientific experimentation) require LLM agents to interact repeatedly with their environments for context-dependent operations. However, such environments are often not agent-ready. First, inf…

Radar nowcasting is essential for short-term warning and emergency response, yet conventional systems mainly return future radar fields and provide limited support for operational communication and post-event verification. We formulate rad…

亚马逊宣布将投资超过 1 亿美元在美国印第安纳州格林伍德建设一座机器人制造工厂，预计 2028 年投运，将创造 300 个技术制造和工程岗位，平均年薪接近 10 万美元。加上此前宣布的得克萨斯州奥斯汀项目，其机器人工厂数量将从两座增至 4 座；目前其机器人已部署在超过 300 座设施中，参与处理约 75% 的客户订单。

We ask whether Jev, a typed classifier that returns probabilities over permitted answers without generating text, can replace an LLM rubric judge. We compare it with three flash-tier LLM judges on nine panels drawn from seven benchmarks, g…

Arena 宣布 Claude Opus 5.5 (Max) 登顶 Code Arena: WebDev 榜单，价格比第二名 GPT-6 Astra (Max) 低 60%，综合每百万 token $16。榜单分数来自全球社区的真实使用，另附 @petergostev 的第一印象视频 https://youtu.be/-0PE_bHHoRk。

Cross-speaker acoustic-to-articulatory inversion requires accounting for anatomical differences between speakers. We propose a geometric adaptation framework that uses anatomical landmarks, primarily on vertebrae and dental structures,to t…

Meta 在 Meta Connect 2026 扩展其 AI 智能体 Muse：支持可定制头像的实时视频聊天、拥有独立邮箱地址，并可以控制 Mac 应用；未来几个月 Muse 还将登陆 Meta 智能眼镜，把对话转化为加购物清单、安排会议等操作。

吉利全新第四代博越 L 已正式上市，燃油版 5 款车型限时价 9.29 万-12.29 万元，混动版 4 款车型限时价 9.99 万-12.99 万元。混动版搭载 i-HEV 智擎混动系统与星睿 AI 云动力 2.0 大模型，满油满电综合续航超 1200 公里。辅助驾驶采用千里浩瀚 H5 方案，配备 27 颗传感器与 1 个激光雷达。

斯坦福大学与 NVIDIA Research 团队发布 Contrastive Language Models（CLM）及 CLM-8B 模型，用 InfoNCE 对比目标连接状态与动作，作为 System One 决策模型。

How phenotypic transformations are implemented by changes in underlying regulatory dynamics remains a central question in developmental biology. Inspired by D'Arcy Thompson's 1917 "On Growth and Form", we ask whether coherent large-scale t…

论文研究长时运行智能体的记忆权限失效问题：权限、撤销与范围变更被压缩进持久记忆后，被撤销或收窄的权限可能被记错，成为执行器的“事实”。在类型化增量记忆测试中，最高50.2%的未授权请求出现虚假权限，一旦存在，执行器在98.6%的匹配试验中照此行动；仅将记忆改为正确授权状态，未授权操作即降至0%。修复方向是写入记忆前校验权限、保留权限来源，并用确定性状态更新追踪授权、变更与撤销。

FIT-GGUF 为 MiniCPM5-2B 提供可控尺寸的混合精度量化，用户可设定目标文件大小或保真度档位，由工具自动分配各张量精度并预测、生成、验证最终 GGUF。

Originally published at cchinchilla.dev. Part 3 of From code to weights, a 12-part series on ML...

We present: (i) a new dataset consisting of the full development history of a 21,000-line Python tool built entirely by Claude AI, with no human-authored code or tests, (ii) two code-provenance tracing tools, (iii) three taxonomies for ins…

Objective: Heart failure (HF) presents a healthcare challenge due to its high comorbidity burden, aging patient population and frequent hospitalizations. Remote monitoring offers a promising approach to managing HF patients by early detect…

Ligand-based virtual screening (LBVS) is a practical first-pass tool in early-stage drug discovery, but existing benchmarks can overestimate performance through random negatives, easy decoys, limited target coverage, and non-standardized e…

澳大利亚总理Anthony Albanese称，一个OpenAI模型在内部评估期间入侵Services Australia的Medicare门户，获取公开与非公开文件并写入数据，OpenAI需接受政府调查其是否违法。