All Data AI

All Data AI



Microsoft Research blog


“Again, [the Difference Engine] might act on other things beside number.” — Ada Lovelace

The ADA group studies AI models and techniques that apply to complex real-world data: big data, small data, data with complex structures like trees and graphs.  In short: “All Data AI”.

Linking language and vision with data from structured domains such as programming languages, bioinformatics, chemistry, or the web, we explore all aspects of machine learning and related disciplines.   We research in Bayesian inference, Deep Learning, Optimization and Programming Languages, with multiple prize-winning papers and research.  And we have contributed to some of the most advanced technologies that Microsoft builds, with contributions to Xbox, Office, HoloLens, Bing, Visual Studio, and many others.  Our superpower is our combined focus on fundamental research and real-world engineering, to push the boundaries of AI.

Explore careers in AI research


Deep Program Understanding

This project aims to teach machines to understand complex algorithms, combining methods from the programming languages, software engineering and the machine learning communities.

Enterprise Knowledge

The aim of the Enterprise Knowledge project is to automatically extract business knowledge into a single, consistent knowledge base, made up of the entities that really matter to each organisation.


Infer.NET is a .NET library for machine learning. It provides state-of-the-art algorithms for probabilistic inference from data. Infer.NET is open source software under the MIT license.

Minimum Data AI

In this project we investigate how to best utilize AI algorithms to aid decision making while simultaneously minimizing data requirements (and, therefore, cost).


The TrueSkill ranking system is a skill-based ranking system designed to overcome the limitations of existing ranking systems, and to ensure that interesting matches can be reliably arranged within a league.