Crescent helps make digital IDs private by preventing tracking across uses while letting users only disclose what’s necessary from their credentials: https://msft.it/6044snCgM
Microsoft Research
Think Tanks
Redmond, Washington 326,708 followers
We advance science and technology to benefit humanity.
About us
At Microsoft Research, we accelerate scientific discovery and technology innovation to empower every person and organization on the planet to achieve more. We do this by bringing together the best minds across diverse disciplines and backgrounds to take on the most pressing research challenges for Microsoft and for society. Our Research Lens We consider research directions through the lens of the positive impact we aspire to create with and for customers, communities, and all of society.
- Website
-
http://www.microsoft.com/research
External link for Microsoft Research
- Industry
- Think Tanks
- Company size
- 1,001-5,000 employees
- Headquarters
- Redmond, Washington
- Founded
- 1991
Updates
-
TimeCraft introduces an open-source framework for generating synthetic time-series data through few-shot samples, natural language prompts, and model feedback; Velocitune dynamically reweights domains in continual pretraining to improve reasoning and efficiency; new research examines how practitioners measure representational harms in LLMs and how tools can better meet their needs; the SPACE of AI study reveals how AI tools are shaping developer productivity and satisfaction across teams; and TACO-RL uses reinforcement learning to compress prompts in a task-aware way, cutting latency and inference costs while improving accuracy.
-
Each Microsoft Academic Research Exchange session highlights real-world research applications, interdisciplinary collaboration, and emerging methodologies. Register here for the upcoming webinar: https://msft.it/6041sn7oR
-
-
For the finale of “The AI Revolution in Medicine, Revisited,” Peter Lee, Carey Goldberg, and Dr. Isaac Kohane compare their predictions to insights from the series’ most recent guests, including leaders in AI-driven medicine and doctors in training. https://msft.it/6045snnk5
-
-
Microsoft Research reposted this
Super excited about Cosmos AI and CoMET coming out today: https://lnkd.in/gCF29AMB. Really grateful for the opportunity to contribute to this joint work by Epic, Microsoft Research, Yale School of Medicine, as featured by Seth Hain and Phil Lindemann at Epic #UGM2025. Our research collaboration with Seth Hain, Rahul Shah, and team originated two years ago, sparked by the GenAI revolution and our shared excitement about the generational opportunity to improve patient care by leaps and bounds. It has gone a long way since! The confluence of biomedical digitization and GenAI makes the long-standing dreams of real-world evidence and precision health quickly become a reality. In particular, generative medical event models have emerged as a promising direction for pretraining high-fidelity virtual patients that can serve as patient digital twins for prognosis and predictive medicine. Just like GPT-4 learns from the web by predicting the next tokens, a generative medical event model can learn from real-world patient journeys by predicting the next medical events. When pretrained at a sufficiently large scale, such a foundation model can potentially acquire "emergent capabilities" by distilling hereto unknown population-scale insight from real-world data. There is exciting work along this direction, such as CLMBR, MOTOR, Foresight, ETHOS. At Microsoft Research, we conducted the first study on scaling laws for structured patient journeys: https://lnkd.in/gHb9FQjq. The results unveiled interesting commonalities and differences compared to natural-language text. However, these prior works, ours included, are inherently bounded by the limited scale of publicly available data. Enter #Cosmos, a unique population-scale real-world evidence platform created by the community of health systems using Epic. So wonderful to see our research collaboration yield the first significant milestone, where our joint team has pretrained the first truly population-scale patient foundation model (CoMET) and conducted the largest study to this date on scaling laws for structured patient journeys, with promising results across 78 real-world tasks. Personally, it's simply a privilege to work with such a talented and hard-working team singularly committed to improving human health. I still remember vividly the night of Black Friday in 2024, when Shane Waxler pinged me on Teams to share the first successful pretraining run on #Cosmos :) Congrats to Seth Hain, Rahul Shah, and all the co-authors: Shane Waxler, Paul Blazek, Davis White, Daniel Sneider, Kevin Chung, Mani Nagarathnam, Patrick Williams, Hank Voeller, Karen Wong, Matthew Swanhorst, Sheng Zhang, Naoto Usuyama, Cliff Wong, Tristan Naumann, Andrew Loza, MD, PhD, Daniella Meeker! Ofc, we're still far away from the moonshot of attaining precision health and democratizing the best healthcare for everyone in the globe. So much remains to be done!
-
-
MindJourney can enable AI to navigate and interpret 3D environments from limited visual input, potentially improving performance in navigation, planning, and safety-critical tasks: https://msft.it/6040spq8u
-
-
In Ep. 10 of “The AI Revolution in Medicine, Revisited,” former Washington State Secretary of Health Dr. Umair A. Shah, M.D., M.P.H. & Mayo Clinic CEO Dr. Gianrico Farrugia, M.D. share their optimism for how AI can improve public health, healthcare delivery, & medical research. https://msft.it/6049sRd0V
-
-
Microsoft Research reposted this
Featured on the cover of Science Magazine, BioEmu from Microsoft Research, a deep learning system that rapidly generates diverse protein conformations for accurate insights into protein function. Illustration: N. Burgess/Science; Data: S. Lewis et al., https://lnkd.in/gev6H7Dw
-
-
What can domains like cybersecurity and pharma teach us about governing AI and the role of evaluation and testing? Find out in the “AI Testing and Evaluation” podcast and in a new white paper distilling insights from eight expert case studies. https://msft.it/6047suIdJ
-
RetroChimera, now available on Azure AI Foundry, marks a new milestone for predicting synthesis routes to drug-like molecules, opening new possibilities for AI in drug discovery. Learn more: https://msft.it/6040sutzy
-