Loading...
无界
Articles

微软亚洲研究院新著问世:《无界——透视微软创新研究之境》 

May 15, 2025

以人工智能大模型的崛起为标志,计算机科学踏入了一望无垠的未知之境。我们对技术变革带来的无限可能满怀期待,却也在颠覆性的重塑中遭遇诸多困惑与挑战: 新的研究范式已然到来,如何发现并投身于具有持久影响力的研究方向?面对无先例可循的技术“无人区”,突破性创新需要怎样的思维破壁?AI时代,需要什么样的人才?什么样的成长环境能够激发人才的潜能?企业如何构建持续创新的内生动力?前沿成果又该以何种方式、何种形态...

background pattern
Articles

从直觉到“深度思考”:多维进化的大模型推理能力 

May 14, 2025

编者按:尽管近年来人工智能的能力迅速增强,但在复杂的推理任务中仍存在不足。微软亚洲研究院的研究员们从多个角度对此展开研究,不断探索提升大模型推理能力的新途径。从利用蒙特卡洛树搜索模拟人类“深度思考”过程的 rStar-Math,到基于规则的强化学习方法 Logic-RL;从融合大语言模型数学直觉与符号方法的 LIPS,到提升自动形式化准确性的新框架;再到自动生成高质量、有监督数学数据的神经符号框架...

EpiCoder
Articles

特征树驱动的数据合成框架,加速构建高质量模型基座 

May 14, 2025

作者:工程与基础架构组 编者按:随着代码大模型能力的不断增强,高质量指令数据的构造成为释放其潜力的关键。然而,现有方法普遍依赖单一的代码片段作为构造种子,限制了数据的复杂度与多样性。近日,微软亚洲研究院联合厦门大学、清华大学提出全新特征树驱动的数据合成框架,通过建模代码语义层级关系,实现了对合成代码复杂度的精细控制,并支持从函数级到多文件级的多样任务生成。基于该框架训练得到的 EpiCoder 模...

diagram
Articles

TimeDP: Creating cross-domain synthetic time-series data 

May 13, 2025

Time-series data—measurements collected over time like stock prices or heart rates—plays a vital role in AI forecasting systems across industries. As these systems advance, the need for time-series data is increasing, especially synthetic data, which offers numerous advantages over real-world…

text
Articles

Teaching LLMs to think: Xian Zhang on advancing mathematical reasoning in AI 

May 13, 2025

Math is more than a school subject—it's the engine behind scientific discovery, driving advances in everything from climate modeling to AI. At Microsoft Research Asia, senior researcher Xian Zhang is leading efforts to help AI move beyond surface-level pattern recognition toward…

The image shows a radar chart comparing the performance of different AI models across various metrics. The chart has a circular grid with labeled axes including VO, AS, CEc, CEe, CL, MCr, MCt, MCu, MS, QLI, QLqA, SNs, KNa, KNc, KNF, KNn, and AT. Different AI models are represented by various line styles: Babbage-002 (dotted line), Davinci-002 (dash-dotted line), GPT-3.5-Turbo (dashed line), GPT-4.0 (solid thin line), OpenAI ol-mini (solid thick line), and OpenAI o1 (solid bold line). There is a legend in the bottom left corner explaining the line styles for each model. The background transitions from blue on the left to green on the right.
Microsoft Research Blog

Predicting and explaining AI model performance: A new approach to evaluation 

May 12, 2025 | Lexin Zhou and Xing Xie

ADeLe, a new evaluation method, explains what AI systems are good at—and where they’re likely to fail. By breaking tasks into ability-based requirements, it has the potential to provide a clearer way to evaluate and predict AI model performance.

man standing and looking at a digital scene

In the news | The Microsoft Cloud

Microsoft as customer zero: Empowering research teams with AI 

May 12, 2025

Research has always been an integral part of Microsoft’s identity, driving our role as a global technology leader. Since 1991, Microsoft Research has dedicated itself to a fundamental research approach—advancing knowledge, deepening our understanding of the world, and exploring how…

In the news | Mayo Clinic

Peter Lee, Ph.D., president of Microsoft Research, elected to Mayo Clinic Board of Trustees 

May 9, 2025

Peter Lee, Ph.D., president of Microsoft Research, was elected today to the Mayo Clinic Board of Trustees. Dr. Lee oversees Microsoft Research's 13 global laboratories, driving advancements in artificial intelligence, computer science, health and life sciences while fostering the development…

Illustrated headshots of Hongxia Hao (left) and Bing Lv (right).
Microsoft Research Podcast

Abstracts: Heat Transfer and Deep Learning with Hongxia Hao and Bing Lv 

May 8, 2025 | Gretchen Huizinga, Hongxia Hao, and Bing Lv

Silicon has long borne the burden of heat transfer in electronics, but in a post-Moore’s Law world, researchers like Hongxia Hao and Bing Lv are using AI to discover and design next-generation materials that exceed the limits of silicon’s thermal…

  • Previous
  • 1
  • …
  • 20
  • 21
  • 22
  • 23
  • 24
  • …
  • 568
  • Next