Assisted Algorithms

Aegon: Self-Auditable Key Transparency

KITE: a Kinesthetic Interface for Tangible Exploration

Scalable Training of Diffusion Samplers via Fixed-point Score Matching

AI World Models for Memory, Meaning, and Human Futures

Sparrow Studio Tutorial - Ep 1. - What is Sparrow Studio?

What AI gets wrong and what failure teaches us

Generative Gibbs sampling: Composing generative models with explicit physical context

One-shot generative flows: Existence and obstructions

The Future of the Grid: How AI Transforms Power System Planning

Introducing Quine: An AI research system designed for the complexity of biology

Latch-Free Index Trees That Scale: Avoiding Thread Switches & Stalls

Sample and Map from a Single Convex Potential: Generation using Conjugate Moment Measures

Probabilistic Inference for Controlling Diffusion Models

MSRNE Generative Modeling & Sampling Workshop | Day 2

MSRNE Generative Modeling & Sampling Workshop | Day 1

AI agents for therapeutic reasoning across biological contexts

Analyzing Group Chat Encryption in Messaging Applications

Invited Research Talk: Measuring Generalization in EEG Foundation Models

TruthTable: A Verifiable Query Engine

AI, Radio Astronomy, and the Search for Life Beyond Earth

PhylaFlow: Hybrid flow matching in phylogenetic tree space

Binomial Flows: Denoising and flow matching for discrete ordinal data

Reinforce Adjoint Matching: Scaling Diffusion RL

Rare event analysis via stochastic optimal control

New tools, models, repos, and papers out of Microsoft Research are here.

Guiding the AI disruption to the Good Place

New fine-tuning of language models: Match meaning, not tokens

Introducing Interwhen: Steering reasoning agents with real-time verification

Introducing GitHub Agentic Workflows: AI that runs your repo

MagenticLite: A full-stack agentic experience powered by Small Models

Microsoft Research Forum | Season 2, Episode 4

Storycaster: An AI System for Immersive Room-based Storytelling (CHI 2026)

Can we AI our way to a more sustainable world?

Microsoft Transforms its Cloud Supply Chain with Optimization and Generative AI

Ideas: Steering AI toward the work future we want

Will machines ever be intelligent?

ARO: A new lens on matrix optimization for LLMs

Agent Lightning: One learning system that makes all agents evolve

Lessons from deploying HealthBots with experts-in-the-loop

OptiMind: Teaching small language models to think like optimization experts

Magentic Marketplace: Testing societies of agents at scale

Dion2: A new simple method to shrink matrix in Muon

Microsoft Research Forum | Season 2, Episode 3

Trailer: The Shape of Things to Come

Microsoft Research India 2025 Highlights

Inside Microsoft Research Asia: Fundamental research powering the future of AI

Data Formulator: Vibe with your data, in control

Accelerating MRI image reconstruction with Tyger

Tool-space Interference: An emerging problem for LLM agents

Publicly-verifiable elections

A brain-inspired agentic architecture to improve planning with LLMs

Microsoft Research Forum | Season 2, Episode 2

Ideas: Community building, machine learning, and the future of AI

Coauthor roundtable: Reflecting on healthcare economics, biomedical research, and medical education

Promptions

Data Formulator v0.5: Vibe with data, in control

Ultra Ethernet for next-generation AI and HPC workloads

Evaluating the Cultural Relevance of AI Models and Products: Insights from the YUX Team

Evaluating Privacy Policies under Modern Privacy Laws At Scale: An LLM-Based Automated Approach