AI-agent-based tool for tuning and optimizing system prompts, and particularly those of Microsoft Copilot offerings. It is used to replace manual adaptation of prompts -- requiring extensive human labor -- in scenarios such as when the underlying language models are updated. Microsoft Copilot offerings use…
In a Magician's magical cabin alone in a serene forest, an alien walking on the floor, starting from the cabin's door to the mow near the bottom right corner of this image. Four characters stood on the stage. In front of the stage, a man…
Guidance | Control LM outputs. Reduce latency and cost. Guidance is a proven open-source Python library for controlling outputs of any language model (LM). In only one API call, developers express in Python the precise programmatic constraints the model must follow for structured output in…
A flexible 3D foundation model of the atmosphere. Aurora, developed by a team of Microsoft researchers, is a cutting-edge AI foundation model that can extract valuable insights from vast amounts of atmospheric data. This 1.3 billion parameter model excels at a wide range of prediction…
Bridging technological and community divides by enriching Generative AI research with the rich diversity of African languages. Join us in shaping a more inclusive digital future Microsoft Research Africa is pioneering advancements in Generative AI through our African Languages Research initiative, focusing on advancing Large…
Project AOC is building an analog optical computer that has the potential to accelerate AI inference and hard optimization workloads by 100x. To achieve this, we rely on a physical system to embody the computations and step away from several fundamentally limiting aspects of digital…
Creators, players and AI models collaboratively build and evolve games over time In this project, we explore how game creators, players and generative AI models can collaboratively build and continually evolve games over time. We are currently looking at game narratives and non-player character definitions…
An AI enabled digital health solution to address delays in maternal care seeking While being amongst the most developed African countries, Kenya still has one of the world’s highest maternal mortality rates (51x that of the United Kingdom), with over 6,000 women dying during childbirth each…
Driving large vision models with video demonstrations In-context learning for vision data has been underexplored compared with that in natural language. Previous works studied image in-context learning, urging models to generate a single image guided by demonstrations. In this project, we propose and study video…
Million-Tokens Prompt Inference for Long-context LLMs MInference 1.0 leverages the dynamic sparse nature of LLMs' attention, which exhibits some static patterns, to speed up the pre-filling for long-context LLMs. It first determines offline which sparse pattern each head belongs to, then approximates the sparse index online…
Using AI to Enhance the Way We Think Much AI research focuses on solving specific tasks for people - generating content or automating processes. While such systems may be powerful, there are risks that this approach may bring to human cognition. They may impact the…
Controlling Time-Varying Emotional States of Flow-Matching-Based Zero-Shot Text-to-Speech EmoCtrl-TTS is an emotion-controllable zero-shot TTS that can generate highly emotional speech with non-verbal vocalizations such as laughter and crying for any speaker. EmoCtrl-TTS is purely a research project. Currently, we have no plans to incorporate EmoCtrl-TTS…
Meet our community of researchers, learn about exciting research topics, and grow your network
Ongoing conversations at the cutting edge of research
Join us for a continuous exchange of ideas about research in the era of general AI
