Curating the frontier of Artificial Intelligence

Time-Varying Data as Sheaves: an Invitation to Narratives

现代科学和工程越来越依赖于时变数据,，但用于模拟时间现象的数学工具经常被开发出来......

Modern science and engineering increasingly rely on time-varying data, yet the mathematical tools used to model temporal phenomena are often developed...

Everything in Moderation: Per-Domain Coverage Optima and Alignment-Resistant Domain Gaps in Multi-Domain Mid-Training

训练中期, 预训练和对齐, 之间的阶段是模型的 每个域数据组合通常由数据可用性设置的阶段，而不是...

Mid-training, the stage between pre-training and alignment, is where a model's per-domain data composition is typically set by data availability rathe...

The Surprising Effectiveness of Approximate Value Iteration in Self-Play

将搜索与函数逼近相结合推动了游戏程序的重大进步, 使自玩算法比以往更具竞争力......

Combining search with function approximation has driven major advances in game-playing programs, making self-play algorithms more competitive than eve...

MIT Schwarzman College of Computing launches pilot to help educators teach AI across disciplines

“我们希望让学生成为人工智能的批判性思考者,，而不仅仅是技术的使用者,” Asu Ozdaglar, 学术副院长...

“We want to empower students to become critical thinkers about AI, not just users of the technology,” says Asu Ozdaglar, deputy dean of academics for ...

SAEScientist-Bench: Can AI Agents Conduct Autonomous SAE Interpretability Research?

虽然递归自我改进(RSI)的研究主要采用自动化模型训练管道,，但可靠的自主开发需要错误的...

While research on recursive self-improvement (RSI) has predominantly automated model training pipelines, reliable autonomous development demands a mis...

MeClear: Cooperative Game-Theoretic Attribution and Risk-Aware Memory Clearance for Long-Horizon LLM Agents

长期大型语言模型 (LLM) 代理依靠外部存储系统来跨扩展交互保留用户偏好和任务知识...

Long horizon Large Language Model (LLM) agents rely on external memory systems to preserve user preferences and task knowledge across extended interac...

A Generalization of Amari's Bayesian Duality

Amari 对信息几何和机器学习的贡献是众所周知的。在这里,我们重新审视Amari的关于贝叶斯对偶性的工作，该工作尚未记录...

Amari's contributions to information geometry and machine learning are well known. Here, we revisit Amari's work on Bayesian duality which has not rec...

ExecCritic: Learn to Test, Test to Improve for Coding Agents

执行反馈可以指导编码代理进行正确的存储库修复,，但前提是测试捕获了问题所请求的行为。年龄...

Execution feedback can guide coding agents toward correct repository repairs, but only when the tests capture the behavior requested by the issue. Age...

Co-Evolving Harnesses and Models: On-Policy Correction Helps Weaker Models Catch Up Where Imitation Fails

代理利用(系统提示,工具集,执行挂钩,和围绕模型的上下文管理支架)是代理的关键决定因素...

Agent harnesses (the system prompt, tool set, execution hooks, and context-management scaffolding around a model) are a critical determinant of agenti...

A Data-Driven Framework for Identifying and Prioritizing RPA Opportunities in Healthcare Processes

机器人流程自动化 (RPA) 被广泛用于减轻美国医院的行政负担,，但估计有 30-50% 的 RPA 计划...

Robotic Process Automation (RPA) is widely used to reduce administrative burden in United States hospitals, yet an estimated 30-50% of RPA initiatives...