Loading...
article feature image
Articles

MIRA:让AI真正读懂生命体征的”不规则律动” 

April 7, 2026

编者按:医疗AI在影像诊断与病历理解领域已展现出惊人的潜力,但一个更根本的挑战正浮出水面:模型能够读懂静态的"快照",却难以捕捉生命的"动态流转"。现有方法多依赖规则采样与插值预处理来强行对齐数据,面临着信息失真、噪声引入、临床适应性差等局限,更引发深层质疑——若模型始终需要人工规整才能理解时间,那它是否真正“领悟”了生理演变的内在规律? 对此,微软亚洲研究院推出MIRA(Medical Irre...

article feature image
Articles

通用语音识别模型VibeVoice ASR:长达60分钟音频一次性“直出”结构化转写 

April 6, 2026

语音识别技术在近年来取得了飞速发展,但在面对真实的复杂场景时,许多系统依然显得有些力不从心:漫长的会议、多人混杂的交谈、极具专业性的行业术语,或是中英夹杂的日常对话…… 传统的语音识别系统往往需要将长音频切分成一个个短小的片段,分别识别后再费力地“拼凑”起来。这种“化整为零”的方法不仅容易丢失上下文,还常常在谁说了什么、什么时候说的等问题上张冠李戴。 近日,微软亚洲研究院发布了一款通用的语音识别模...

feature article cover
Articles

AI Next 播客 | 对话李东胜:AI 与人脑,正在彼此“教会”对方什么? 

April 6, 2026

《AI Next》是微软亚洲研究院推出的一档利用 AI 技术制作的播客,内容聚焦 AI 前沿技术、科研趋势与社会影响。第一季主要围绕当今智能发展的核心议题,探索前沿趋势。 在第六期节目中,我们邀请到微软亚洲研究院首席研究员李东胜博士,一同探讨 AI 与大脑的深层关联。以人脑为灵感而诞生的神经网络,究竟与大脑是单纯的模仿与被模仿关系,还是存在更深度的联结?AI 该如何向历经亿万年进化的人脑学习高效的...

feature article cover
Articles

别让AI在部署后停滞:OEL重塑大模型进化之路 

April 6, 2026

大模型部署之后能力还可以持续提升吗? 如果去问AI开发者这个问题,在过去,答案大多是否定的。传统的大模型在出厂那一刻,其智力水平就几乎定型了。即便在实际应用中遇到了未见过的新问题,或者被用户反馈指出错误,模型也很难像人类一样,在下一次尝试时立刻吸取教训。这正是当前大语言模型面临的尴尬现状。 在现有的主流范式下,AI模型性能的提升高度依赖于离线训练,要么是昂贵的人工标注,要么是预先构建的模拟环境。一...

ADeLe | Three white line icons, showing a circle with a checkmark, a search document, and a set of tools, on a blue‑to‑green gradient background.
Microsoft Research Blog

ADeLe: Predicting and explaining AI performance across tasks 

April 1, 2026 | Lexin Zhou and Xing Xie

AI benchmarks report how large language models (LLMs) perform on specific tasks but provide little insight into their underlying capabilities that drive their performance. They do not explain failures or reliably predict outcomes on new tasks. To address this, Microsoft…

AsgardBench | three whit icons on a blue to purple gradient background | first icon shows a laptop screen with a eye in the upper right corner, second icon shows relational nodes | third icon is a security shield with a checkmark
Microsoft Research Blog

AsgardBench: A benchmark for visually grounded interactive planning 

March 26, 2026 | Andrea Tupini, Lars Liden, Reuben Tan, Yu Wang, and Jianfeng Gao

Imagine a robot tasked with cleaning a kitchen. It needs to observe its environment, decide what to do, and adjust when things don't go as expected, for example, when the mug it was tasked to wash is already clean, or…

V2GP framework | Three white line icons, showing a target within a rounded square, a checklist, and a robotic arm, on a blue‑to‑green gradient background.
Microsoft Research Blog

GroundedPlanBench: Spatially grounded long-horizon task planning for robot manipulation 

March 26, 2026 | Sehun Jung, HyunJee Song, Dong-Hee Kim, Reuben Tan, Jianfeng Gao, Yong Jae Lee, and Donghyun Kim

Vision-language models (VLMs) use images and text to plan robot actions, but they still struggle to decide what actions to take and where to take them. Most systems split these decisions into two steps: a VLM generates a plan in…

The Shape of Things to Come podcast | illustration of Nicolo Fusi, Doug Burger, and Subutai Ahmad
Microsoft Research Podcast

Will machines ever be intelligent?  

March 23, 2026 | Doug Burger, Subutai Ahmad, and Nicolo Fusi

Are machines truly intelligent? AI researchers Subutai Ahmad and Nicolò Fusi join Doug Burger to compare transformer-based AI with the human brain, exploring continual learning, efficiency, and whether today’s models are on a path toward human intelligence.

article feature image
Articles

微软亚洲研究院携手 CCF,共促科研创新生态 

March 16, 2026

近日,2025 CCF 颁奖典礼在北京举行。大会以“责任 · 创新 · 奉献”为主题,集中表彰了在计算技术研究、学术服务与科研生态建设方面作出积极贡献的个人与团队。作为大会的协办方之一,微软亚洲研究院参与了多项重要环节,与学术界同仁共同见证了这一年度科研盛会。 在本届大会的个人奖项中,微软研究院学术合作总监马歆女士荣获 CCF 年度志愿者(服务类)奖。在过去十余年中,马歆积极促进微软亚洲研究院与C...

  • Previous
  • 1
  • 2
  • 3
  • 4
  • …
  • 576
  • Next