Loading...
chart
Articles

微软亚洲研究院发起全球人工智能价值观挑战赛,共建AI的伦理基石 

May 28, 2026

我们习惯了向AI提问宇宙的奥秘、代码的逻辑、生活的常识,但很少有人追问:当诚实与善良发生冲突,AI该如何选择?在缺少相关数据和信息的情境下,AI是否能做出符合人类普世价值观与伦理道德的选择?这些问题并非抽象的哲学思辨,而是AI走向社会、成为融入我们日常生活必须面对的真实考题。 为探索AI在复杂现实中对齐人类价值观与伦理的核心能力,微软亚洲研究院正式发起“全球人工智能价值观挑战赛(Global AI...

Cover image
Articles

AI解码脑疾病,开启精准诊疗新可能 

May 28, 2026

编者按:阿尔茨海默病、帕金森病等神经退行性疾病,长期面临早期难诊断、药物难入脑、病程难预测等挑战,困扰着无数患者与家庭。如今,人工智能的跨越式发展,正在为这一领域带来新突破。 微软亚洲研究院(上海)深耕人工智能与脑科学交叉领域,以三大创新推动神经疾病诊疗升级:通过多模态融合提升阿尔茨海默病早期信号识别能力,以人机协作打通脑靶向药物递送路径,并借助个性化模型实现帕金森病病程轨迹精准预测。 随着全球人...

graphical user interface
Articles

大模型智商在线,为何“情”商掉线? 

May 28, 2026

如今的大语言模型早已在数学领域展现出令人惊叹的实力,在 AIME 美国数学邀请赛、IMO 国际数学奥林匹克竞赛等高难度抽象数学竞赛中,多款主流模型都能交出近乎满分的答卷。这一表现也成为大模型推理能力持续突破的亮眼标签。 但是,这些在抽象数学题中所向披靡的“学霸”,面对藏在现实场景中的数学问题时却频频失手,抽象解题能力始终无法转化为可靠的现实应用能力,形成了巨大的性能落差。比如模型能精准解出抽象的数...

graphical user interface
Articles

ICLR上新 | 多模态生成与具身智能的前沿突破 

May 28, 2026

编者按:欢迎阅读“科研上新”栏目!“科研上新”汇聚了微软亚洲研究院最新的创新成果与科研动态。在这里,你可以快速浏览研究院的亮点资讯,保持对前沿领域的敏锐嗅觉,同时也能找到先进实用的开源工具。 4月23日至4月27日,深度学习与表征学习领域最负盛名的学术会议之一的ICLR,将在巴西里约热内卢举行。我们通过两期“科研上新”为大家带来多篇微软亚洲研究院入选ICLR 2026的精选论文解读。在第一期中,我...

Three icons (speech bubble, handshake, and interconnected circles) on a blue and green gradient background.
Microsoft Research Blog

Extending Human Intelligence Through AI 

May 27, 2026 | Ken Archer and Harald Wiltsche

Understanding AI as an extension of human intelligence—not a replacement for it—offers a more grounded path for building trustworthy AI systems.

graphical user interface, application
Articles

Fara1.5 – A family of frontier computer use agent models 

May 21, 2026

By: Ahmed Awadallah, Sahil Gupta, Yash Lara, Yadong Lu, Hussein Mozannar, Akshay Nambi, Zach Nussbaum, Yash Pandya, Aravind Rajeswaran, Corby Rosset, Alexey Taymanov, Luiz do Valle, Vibhav Vineet, Spencer Whitehead, Andrew Zhao We are excited to introduce the Fara1.5 family…

MagenticLite | Four white line icons—a document in the cloud, a flowchart with three boxes, chart, and a gear—displayed on a teal‑to‑green gradient background with a subtle textured pattern.
Microsoft Research Blog

MagenticLite, MagenticBrain, Fara1.5: An agentic experience optimized for small models 

May 21, 2026 | Microsoft Research AI Frontiers

MagenticLite is an agentic system for small models that works across the browser and local file system in a single workflow. It combines specialized models and orchestration to support efficient agentic performance on everyday tasks.

Three white line-style icons centered on a textured blue-to-green gradient background: a shield, a checkmark inside a circle, and an ID card with a user silhouette—representing security, verification, and digital identity.
Microsoft Research Blog

Vega: Zero-knowledge proofs for digital identity in the age of AI 

May 21, 2026 | Srinath Setty

Vega turns a full credential into a single proof, sharing only what is needed and nothing more, with performance that works in real apps.

Several people stand in a circle with their hands extended toward the center, including multiple prosthetic arms and hands, photographed from below against a blue sky.

In the news | Microsoft Source

Why better AI starts with the people it often misses 

May 19, 2026

Ask an AI tool for a picture of “someone at work,” and you’ll often get a person at a desk, in front of a computer, maybe holding a coffee cup. When people with disabilities appear at all, the images sometimes…

  • Previous
  • 1
  • 2
  • 3
  • 4
  • …
  • 581
  • Next