Advancing the frontier of quantum chemistry by combining deep learning with Density Functional Theory (DFT) to unlock unprecedented accuracy and scalability in electronic structure simulations. Our mission is to enable predictive modeling of laboratory experiments by achieving chemically accurate electronic structure predictions with deep learning…
Using AI to reshape how we design and innovate advanced materials MatterGen is a diffusion model specifically designed for generating stable inorganic materials across the periodic table. Crucially, the model can be fine-tuned to steer the generation towards a broad range of property constraints, including…
Following the sequence and structure revolutions, predicting functionally relevant protein structure changes at scale remains an outstanding challenge. BioEmu is a deep learning system that emulates protein equilibrium ensembles by generating thousands of statistically independent structures per hour on a single GPU. BioEmu integrates over…
Bringing scale, speed, and precision to molecular discovery with AI We work on accelerating the discovery of small molecules through AI at all steps of the Design-Make-Test cycle. This includes models to predict molecular properties, generative models to design new molecules with the required properties,…
Existing methods for evaluating the factuality of large language model (LLM) responses treat all claims as equally important. This results in misleading evaluations when vital information is missing or incorrect as it receives the same weight as peripheral details, raising the question: how can we…
Real-world data analysis tasks often come with under-specified goals and unclean data. User interaction is necessary to understand and disambiguate a user’s intent, and hence, essential to solving these complex tasks. Existing benchmarks for evaluating LLMs on data analysis tasks do not capture these complexities…
The Psychological Influences of AI (Psi) project explores how AI—especially generative AI—affects human psychology, including cognition, emotion, behavior, and well-being. It investigates topics such as trust, dependency, attention, agency, and the psychological impacts of anthropomorphic AI design.
How do we ensure that AI systems are not only technologically advanced but also deeply resonant and beneficial for the specific communities they aim to serve? As AI becomes more integrated into critical sectors like healthcare, this question has never been more important. This session…
Website privacy policies detail an online service's information practices, including how they handle user data and rights. For many sites, these disclosures are now necessitated by a growing set of privacy regulations, such as GDPR and multiple US state laws, offering visibility into privacy practices…
In this paper we introduce Hierarchical Diffusion Language Models (HDLM) -- a novel family of discrete diffusion models for language modeling. HDLM builds on a hierarchical vocabulary where low-level tokens with detailed semantics are surjectively mapped to high-level tokens with coarse-grained meanings. In the forward…
Microsoft is proud to be a sponsor of the Conference on Language Models (COLM (opens in new tab)), an academic venue dedicated to advancing the study of language modeling. COLM brings together researchers from diverse disciplines to understand, improve, and critically examine the development of…
Recent years have seen rising interest in expanding and equalizing access to low-latency algorithmic trading for a wider group of participants, including retail traders. Cloud-hosted financial exchanges promise a cost-effective platform more accessible to traders. Unfortunately, achieving fairness in cloud environments is challenging due to…
