Curating the frontier of Artificial Intelligence

LongSeeker: Elastic Context Orchestration for Long-Horizon Search Agents

长视野搜索代理必须管理快速增长的工作环境，因为它们推理,调用工具,并观察信息。天真地积累着一切……

Long-horizon search agents must manage a rapidly growing working context as they reason, call tools, and observe information. Naively accumulating all...

Study: Firms often use automation to control certain workers’ wages

当我们听到自动化和人工智能取代工作, 时，似乎一场技术海啸将大量消灭工人……

When we hear about automation and artificial intelligence replacing jobs, it may seem like a tsunami of technology is going to wipe out workers broadl...

Redefining AI Red Teaming in the Agentic Era: From Weeks to Hours

人工智能系统正在进入医疗保健,、金融,和国防,等关键领域，但仍然容易受到对抗性攻击。虽然人工智能红队是......

AI systems are entering critical domains like healthcare, finance, and defense, yet remain vulnerable to adversarial attacks. While AI red teaming is ...

OpenSeeker-v2: Pushing the Limits of Search Agents with Informative and High-Difficulty Trajectories

深度搜索能力已成为前沿大型语言模型(LLM)智能体,不可或缺的能力，但其发展仍然占据主导地位...

Deep search capabilities have become an indispensable competency for frontier Large Language Model (LLM) agents, yet their development remains dominat...

Games people — and machines — play: Untangling strategic reasoning to advance AI

加布里埃莱·法里纳 (Gabriele Farina) 在意大利北部丘陵酿酒区的一个小镇长大。他的父母都没有大学学位,，尽管......

Gabriele Farina grew up in a small town in a hilly winemaking region of northern Italy. Neither of his parents had college degrees, and although both ...

Role-Aware Artificial Intelligence Across Augmentation and Automation in Human-Machine Symbiosis

人工智能(AI)的进化使得人类和计算机器之间的界限变得越来越模糊。在项目中...

The evolution of artificial intelligence (AI) has rendered the boundary between humanity and computational machinery increasingly ambiguous. In the pr...

Instance-Aware Parameter Configuration in Bilevel Late Acceptance Hill Climbing for the Electric Capacitated Vehicle Routing Problem

组合优化中的算法性能对参数设置高度敏感,，而单个全局调整的配置经常失败......

Algorithm performance in combinatorial optimization is highly sensitive to parameter settings, while a single globally tuned configuration often fails...

Learn where to Click from Yourself: On-Policy Self-Distillation for GUI Grounding

图形用户界面(GUI)基础将自然语言指令映射到目标元素的视觉坐标，并作为核心功能...

Graphical User Interface (GUI) grounding maps natural language instructions to the visual coordinates of target elements and serves as a core capabili...

To Call or Not to Call: A Framework to Assess and Optimize LLM Tool Calling

代理人工智能架构通过外部工具, 增强了法学硕士，释放了强大的功能，但可能会产生大量成本。而且,工具还...

Agentic AI architectures augment LLMs with external tools, unlocking strong capabilities but potentially incurring substantial costs. Moreover, tool u...

Position: agentic AI orchestration should be Bayes-consistent

法学硕士擅长预测任务和复杂推理任务,，但许多高价值部署依赖于不确定性, 下的决策，例如, 哪些...

LLMs excel at predictive tasks and complex reasoning tasks, but many high-value deployments rely on decisions under uncertainty, for example, which to...