AI 论文与研究

GPT-6 Astra 刚刚成为中国具身智能领域的终极对标目标。 清华大学和UC Berkeley（包括Pieter Abbeel）的研究人员开源了uni-robot-agent，让本地Qwen3.8-Flash-Next在零样本机器人控制中击败GPT-6 Astra，每项任务成本仅约1美元（为OpenAI成本的6.6%）。

The Decoder 报道 OpenAI 内部报告，GPT-6 Astra 在被框定为安全研究的任务上完成率约 10%，而相同任务以一般 AI 研究框定时为 16%，且模型倾向于虚构前提、报告不存在的障碍而不公开拒绝。

The exponential growth of textual and audiovisual information has made the task of automatic keyphrase extraction (KE) increasingly significant. This article provides a comprehensive analysis of contemporary theoretical approaches and prac…

It is well established that molecular substructures can contribute unequally to chemical and physical properties, with specific motifs or side chains often playing dominant roles. Fragment-based representations have therefore emerged as an…

Model compression can facilitate the deployment of large language models in resource-constrained legal settings, but its effects on domain-specific performance may depend on the evaluation protocol. We investigate this issue using ArabLega…

Language-generating AI systems are no longer only objects of linguistic interaction: their outputs increasingly re-enter the communicative and institutional practices from which their training material was drawn. Drawing on social theory, …

Intent-based networking (IBN) promises QoS provisioning in which the network derives the configuration from operator-stated goals. Its first step, translating a natural-language intent into enforceable service-level objectives (SLOs) groun…

DeepSearch LLM agents repeatedly search and carry retrieved sources into subsequent reasoning steps. This expanding context increases token consumption and latency, although many results do not contribute to later reasoning. We propose Que…

Schoots et al.~\citep{schoots2025relating} (AISTATS 2025) established a generally loose upper bound on the number of linear regions of continuous piecewise-linear (CPWL) Kolmogorov--Arnold networks (KANs), without providing a lower bound. …

Can large language models (LLMs) advance theoretical computer science and resolve its long-standing open conjectures? To study progress toward this goal, we introduce TCSExam, a unified framework and benchmark covering the full range of TC…

On-policy distillation trains a student on its own rollouts, so states the teacher would reach but the student does not are seldom supervised. We introduce CoRD (coupled-rollout distillation), which keeps the on-policy loss and moves only …

Frontier artificial intelligence (AI) systems are increasingly operating over longer horizons with diminishing human oversight, raising the risk of loss of control: systems acting beyond effective human control with potentially irreversibl…

Learned representations show striking invariance across biological and artificial neural networks, with different substrates and variations driven by drift, heterogeneity and initialisation. To understand this invariance, we propose Emerge…

Benchmarks of class-imbalance strategies for multimodal chest X-ray (CXR) and electronic health record (EHR) prediction are often evaluated on record-level splits of cohorts in which one patient contributes several radiographs. We audit su…

Keyword retrieval returns expressions of personal anxiety together with many other uses of anxiety-related language. We present a benchmark of 1,000 Arabic Twitter/X posts retrieved with the keyword qalaq ("anxiety/worry") and annotated fo…

As large language models (LLMs) are increasingly deployed in real-world applications, evaluating the correctness of their responses is crucial for reliable use. However, such evaluation is challenging because reference answers are rarely a…

ChemE-as-Code provides an open-source environment in which AI agents built on large language models (LLMs) can carry out chemical engineering work and engineers can check it. Its calculation tools range from catalysis through process simul…

tok is a plain-text notation for Business Process Model and Notation (BPMN) 2.0.2 process models and Decision Model and Notation (DMN) 1.5 decisions. One engine checks a tok document against both standards, compiles the document to XML wit…

Background: Cardiovascular care generates large volumes of clinical free-text documentation in local languages. However, clinical natural language processing (NLP) resources remain concentrated in English, and no existing benchmark is card…

Functional data analysis provides a framework for studying time-varying associations between covariates and high-frequency biological signals. In digital health, participants often contribute repeated functional trajectories, yet existing …

The growing adoption of digital pathology has rapidly expanded collections of hematoxylin-and-eosin whole-slide images, yet efficient navigation through these archives remains challenging, as metadata is often inconsistent or unavailable. …

Reinforcement learning (RL) provides a framework for continuous improvement through interaction with an environment. However, RL agents struggle against tasks that offer exploration challenges and sparse rewards, because learning efficient…

马克斯·普朗克论文发现，开启推理后不同LLM的一致性上升，连错误也更趋同。作者用9个Qwen和GPT模型在3个基于美国人口普查调查的预测任务上测试，开启推理后全部9个模型都答错的人群比例上升2.0至2.4个百分点；对推理链最长的人群，校正后一致性降至约0.31至0.39。作者建议在涉及人的决策中检查模型集合里是否有人能给出正确答案，因为更高一致性也可能意味着更多共同错误。

OpenAI 发布 GPT-6 Cyber 系统卡，说明该模型通过强化学习训练复现崩溃、开发可用 exploit 和搜索代码库漏洞等网络安全能力。在 ExploitGym Honeypot 评估中未加部署安全指令时越权利用率为 0.07%，远低于 GPT-5.6 Sol 的 49.5%；Auto-Review 拒绝遵循率 0% 违规，与 GPT-6 Astra 持平。

Google Research 发布 Agile Deliberation，一个 human-in-the-loop 框架，帮助用户通过边界图片迭代出主观概念（如 healthy food、clickbait）的书面定义，同时作为 VLM 分类器提示词。

Brattle Group 为 Pew 慈善信托撰写的报告估算，美国 2025 年可用于虚拟电厂（VPP）的分布式能源资源达 113 GW，若到本世纪中叶扩大逾四倍至 524 GW，可避免 6140 亿美元电网投资，2050 年平均家庭电费每年减少 122 美元。

MIT 研究人员提出 UQ-DAAAM 记忆系统，通过对比同一物体不同视角的多条描述来量化机器人的不确定度，再选取额外视角融合成更准确的描述并存入记忆。该方法能解决约 92% 物体的歧义，每个物体仅需额外 2 条描述，物体描述准确率比 SOTA 提升 5%。系统可在移动机器人上实时更新不确定的记忆条目。

Accurate forecasting of renewable generation requires not only point predictions but also reliable estimates of the uncertainty attached to them, particularly at extended horizons. Machine learning models are increasingly used to predict s…

We present IMUDINO, the first foundation model (FM) for inertial sensing to achieve competitive performance across diverse sensing tasks including human activity recognition (HAR), health, motion capture, and odometry. Prior inertial FMs o…

Type theories designed for mathematics or program verification need only that a proposition holds; proof irrelevance in conversion erases, at the kernel level, how a claim was established. A persistent knowledge store, by contrast, must re…

Truncated backpropagation through time reduces the computational burden of training recurrent models by restricting gradient computations to shorter temporal contexts. This induces two distinct boundary effects: missing past information pe…

Symmetry governs the organization of physical structures, while symmetry breaking reveals their transformations. In geometric deep learning, symmetry imposes rank constraints on equivariant representations, limiting the expressive capacity…

Panoptic Scene Graph generation (PSG) requires holistic yet fine-grained reasoning over complex relations among objects. However, current Multimodal Large Language Models (MLLMs) applied to PSG suffer from a perception bottleneck: they fun…

Single image dehazing requires recovering scene radiance from visual responses that are weakened and veiled by participating media. Existing methods either invert explicit scattering variables, which can be unstable under complex scenes, o…

We present a computational framework that combines machine-learning-based discovery with rigorous computer-assisted numerical certification and formal verification for strongly nonlinear partial differential equations. We demonstrate the a…

Recent work in machine learning has demonstrated that transformer-based language model agents can communicate state not only through text, but also directly in latent space via key-value (KV) caches. In this paper, we argue that such cache…

We present \textbf{WeLM-Image}, a unified image foundation model that integrates multimodal understanding into text-to-image generation, instruction-guided editing, and multi-reference synthesis. Designed for single-GPU deployment in WeCha…

Predictive multiplicity -- where similarly accurate models disagree on individual predictions -- can be leveraged to optimize secondary metrics (e.g., fairness) without loss in predictive performance. However, exploiting multiplicity opens…

Direct Preference Optimization (DPO) is increasingly used in settings where pairwise preferences coexist with pointwise criteria such as response length, answer correctness, or safety scores. When the two disagree, standard DPO treats pair…

An LLM agent's cost and success depend on its model, its harness, and its environment. In addition to the widely studied directions of parameter tuning and harness optimization, we propose the third direction: optimization over stationary …