News & features

Research Focus: Week of April 7, 2025
In this issue: We introduce a new dataset designed to assist renewable energy infrastructure planners, a new method for denoising MRI imagery, and an AI tool for analyzing distant galaxies. Check out our latest research and other updates.

A key challenge, and opportunity, of large language models (LLMs) is bridging the gap between their training data and the vast amount of unfamiliar information they encounter in real-world applications. Successfully navigating this divide could unlock a new era of…

VidTok introduces compact, efficient tokenization to enhance AI video processing
| Tianyu He, Junliang Guo, and Jiang Bian
The VidTok method can enable AI systems to process and generate videos more effectively. Compact tokenization reduces computational costs while maintaining video quality across a potentially diverse set of applications.

In the news | The National University of Singapore
NUS collaborates with Microsoft Research Asia to advance AI research and cultivate computing talent

Research Focus: Week of March 24, 2025
In this issue, we examine a new conversation segmentation method that delivers more coherent and personalized agent conversation, and we review efforts to improve MLLMs’ understanding of geologic maps. Check out the latest research and other updates.

Astronomy, born from humanity’s innate curiosity about the stars, has long been a catalyst for revolutionary discoveries. As AI technology advances, intelligent agents powered by large language models (LLMs) are opening up opportunities for astronomers to explore the universe more…

Metasurface: Unlocking the future of wireless sensing and communication
| Lili Qiu and Hao Pan
Metasurfaces explore engineered 2D materials that manipulate electromagnetic and mechanical waves, offering advances in wireless tech. They can power indoor GPS, extend 5G/6G coverage, and enable wireless sensing and imaging.

On the world’s vast canvas, cultures from different regions are like unique brushstrokes, each conveying distinctive perspectives and styles. At Microsoft Research Asia, women researchers from diverse backgrounds have experienced the dynamic interplay of these cultures. Driven by curiosity and…
In the news | The Sequence
Making Small Models Great Achieve GPT-o1 Levels in Math Reasoning with Microsoft rStar-Math
rStar-Math is a novel approach that significantly boosts the mathematical reasoning capabilities of small language models (SLMs). This innovative system enables SLMs to achieve performance levels comparable to, and even exceeding, OpenAI’s o1, despite a significantly smaller model size. This…