Curating the frontier of Artificial Intelligence

3 Questions: Neural transparency and the future of AI design

现在，数百万人正在设计自己的个性化人工智能伴侣,，但大多数人几乎不知道这些创作将如何发挥作用......

Millions of people are now designing their own personalized artificial intelligence companions, yet most have little idea how those creations will act...

DualG-MRAG: Decoupling Macro-Reasoning and Micro-Matching for Multimodal Retrieval-Augmented Generation

虽然多模态检索增强一代 (MM-RAG) 已显示出有希望的结果,，但它仍然难以应对复杂的多跳推理任务。存在...

While Multimodal Retrieval-Augmented Generation (MM-RAG) has shown promising results, it still struggles with complex multi-hop reasoning tasks. Exist...

OSReward: Instituting Standardized Evaluation for Cross-Platform Computer-Use Reward Models

使用计算机的代理 (CUA) 正在数字世界中快速发展。 CUA 轨迹记录了代理的 操作, 状态, 和推理。维里...

Computer-using agents (CUAs) are advancing rapidly across the digital world. A CUA trajectory records the agent's actions, states, and reasoning. Veri...

AISPA: User-Centric System Prompt Auditing for Large Language Model Applications

系统提示是开发人员配置的指令，用于管理人工智能应用程序中基础模型的行为。它们在整个公司中使用...

System prompts are instructions configured by developers to govern the behaviors of foundation models in AI applications. They are used throughout com...

Connecting research to policy on Capitol Hill

今年春天, 25 名麻省理工学院的学生和博士后前往华盛顿与国会工作人员会面，并倡导联邦政府持续投资......

This spring, 25 MIT students and postdocs traveled to Washington to meet with congressional staffers and advocate for sustained federal investment in ...

Connecting research to policy on Capitol Hill

今年春天, 25 名麻省理工学院的学生和博士后前往华盛顿与国会工作人员会面，并倡导联邦政府持续投资......

This spring, 25 MIT students and postdocs traveled to Washington to meet with congressional staffers and advocate for sustained federal investment in ...

Daniela Rus receives Bavarian Minister

该奖项由巴伐利亚州政府和巴伐利亚科学与人文学院, 联合颁发，是科技界最受资助的奖项...

Awarded jointly by the Bavarian State Government and the Bavarian Academy of Sciences and Humanities, it is the most highly endowed award for technolo...

How a medical database developed at MIT evolved into a global standard of data

在通过云, 推动科学数据存储和协作之前，寻求健康研究突破的医学研究人员必须......

Before the advancement of scientific data storage and collaboration via the cloud, medical investigators seeking health research breakthroughs had to ...

Reason-Mediated Behavioral Models for Auditing LLM Social Simulators

大型语言模型越来越多地用作社交模拟器,，包括用作综合调查受访者。大多数评估都会询问是否模拟...

Large language models are increasingly used as social simulators, including as synthetic survey respondents. Most evaluations ask whether simulated ou...

Eviction as Estimation: A Fixed-Lag Smoothing View of Test-Time Memory, and When Measuring Beats Accumulating

具有有限工作记忆的语言模型必须反复决定保留哪些存储项。每个部署的方法都会决定项目到达的那一刻...

A language model with a bounded working memory must repeatedly decide which stored items to keep. Every deployed method decides the moment an item arr...