←
Microsoft Research10,219 posts
Microsoft Research avatar

Microsoft Research

@MSFTResearch

We advance science and technology to benefit humanity.

◷ Joined February 2009▧ 2,553 Photos and videos
Microsoft Research avatar
Microsoft Research@MSFTResearch·18h

GraphRAG Zero solves a fundamental challenge that has limited AI agents' ability to reach their full potential. It’s a technology for holistic discovery that reflects years of Microsoft research, transforming graph-based retrieval from an experimental technique into a deployable system. Optimized for real-world applications and agents, it enables organizations to explore connections across vast amounts of their own data, quickly and at low cost. It’s available now in Foundry IQ and @github’s Repo Mind Light, supporting applications from scientific discovery to repository understanding. msft.it/6017xE5xL

1:10
Microsoft Research avatar
Microsoft Research@MSFTResearch·Nov 16

AI coding agents can waste time and tokens relearning the same lessons. TokenEconomizer builds on past sessions to cut costs while maintaining task success. It’s a path toward coding agents that learn not only how to solve tasks, but also how to spend computation more effectively. The open-source plugin is available now on @github. msft.it/6017xDtyc

0:08
Microsoft Research avatar
Microsoft Research@MSFTResearch·Nov 12

Science fiction, mathematics, and AI research meet in the latest Microsoft Research Focus. Anson Ho explores these breakthroughs, advances in AI coding efficiency, and more.

5:43
Microsoft Research avatar
Microsoft Research@MSFTResearch·Nov 12

Your team just reached a major milestone years in the making. How do you celebrate? Microsoft Research Lab Director Chris White's answer: clean some toilets. Here's why. msft.it/6010x6Se4

0:48
Doug Burger avatar
⟳ Microsoft Research retweeted
Doug Burger@dcburger·Nov 11

When MSR researcher @DimitrisPapail and his collaborators found an important new class of LLM vulnerabilities, they disclosed it to the company, figured out the right mitigations to keep our customers safe, and then disclosed it to the world with the recommended defenses.

Dimitris Papailiopoulos @DimitrisPapail Nov 10 Article Lazare, veni foras: Reprogramming Dormant Neurons to Backdoor AI Agents We find that weight reprogramming fewer than 10 dormant neurons in an LLM is enough to backdoor any agent built on it, with no training required. The model passes evals and behaves normally until a

Microsoft Research avatar
Microsoft Research@MSFTResearch·Nov 11

Cybersecurity isn’t as simple as attackers versus defenders—especially when tech capabilities are rarely single use. Microsoft Research Manager Mike Walker, who leads projects at the intersection of AI and cybersecurity, discusses the state of today’s networked systems and breaks down red, blue, and purple teaming in this podcast conversation with Principal Technical Program Manager Weishung Liu. msft.it/6014x6Gb2

Post attachment
Microsoft Research avatar
Microsoft Research@MSFTResearch·Nov 10

New research from Microsoft Research demonstrates that reprogramming just a few dormant neurons in an AI agent can create a backdoor that leaves the agent behaving normally until it hits a specific trigger. While the specifics of this threat are novel, the mitigation is not: Know where your model came from, and be smart about its permissions. Microsoft Foundry provides critical safety systems for Azure Direct Models and Hugging Face models to ensure that the provenance of the model is secure. Microsoft researcher @DimitrisPapail explains the measures that model hubs and vendors should take as well as mitigation strategies for enterprises and users. msft.it/6014ahp6e

Post attachment
Microsoft Research avatar
Microsoft Research@MSFTResearch·Nov 10

How do you test Copilot against real-world business complexity without exposing customer data? Microsoft researchers created a platform that generates synthetic enterprise environments, complete with simulated employees, meetings, emails, documents, and industry-specific workflows. These environments let Microsoft evaluate Copilot at scale in realistic business scenarios while preserving customer data privacy. Read about the Industry Evaluation Data Generation Platform: msft.it/6011ahGkZ

0:20
Microsoft Research avatar
Microsoft Research@MSFTResearch·Nov 9

Lilika Makabe’s internship provided meaningful connections and opportunities to grow; can AI bring us Star Trek's universal translator; solving a decades-old math challenge; smarter context compaction boosts coding agents; and building a better tokenizer: msft.it/6019aCuhH

2:00
Microsoft Research avatar
Microsoft Research@MSFTResearch·Nov 6

What does it take to grow more confident in your future as a researcher? For Tony Chae, an undergraduate research internship at Microsoft Research offered an opportunity to explore new ideas, collaborate with researchers, and build connections beyond the lab. The experience helped Tony expand his knowledge, develop new perspectives, and build confidence as he prepares for a future PhD.

1:17
Microsoft Research avatar
Microsoft Research@MSFTResearch·Nov 6

Every technological revolution has remodeled our institutions – how might AI? In @ScienceMagazine, Microsoft researcher @glenweyl and co-author @profjamesevans explore ‘reverse alignment,’ the idea that using neural networks as a model for reshaping institutions might offer a path toward harnessing the benefits of AI while managing its risks. msft.it/6015aAHYP

Microsoft Research avatar
Microsoft Research@MSFTResearch·Nov 5

How do you solve the trust problem with AI-generated code? AI agents can generate code at a scale too great for humans to review line-by-line. That makes it difficult to ensure the code will do what the developer intended, and it leaves the final output vulnerable to bugs and security risks. Proof-Driven Development Kit turns plain-language developer intent into a specification in Verus, applying formal methods traditionally used for safety-critical systems. PDD Kit uses mathematical proofs to test agent-generated code against the spec where possible, and clearly indicates where the proofs can’t reach. The result is more reliable agentic code. Researchers @MatthaiP39719 and @alexlavaee break down Proof-Driven Development in the MSR blog: msft.it/6019af2V7

1:25
Microsoft Research avatar
Microsoft Research@MSFTResearch·Nov 3

What if knowing the rules isn't the same as knowing what to do? For Mike Walker, who leads projects at the intersection of AI, cybersecurity, and large distributed systems at Microsoft Research, understanding a system starts with understanding its rules—but expertise also requires something more nuanced: knowing how those rules operate in context. In a new Microsoft Research Podcast episode, Walker joins Principal Technical Program Manager Weishung Liu to talk cybersecurity, hacker culture, autonomous systems, leadership, and what his personal and professional challenges have taught him about judgment and adaptability. First, learn the rules. Then see what's possible. Watch the full conversation: msft.it/6010a4V0m

0:56
Microsoft Research avatar
Microsoft Research@MSFTResearch·Oct 29

Introducing Microsoft Research Focus, a new way to catch up on research news from across Microsoft. In this edition, we're covering advances in AI agents, what patterns in everyday AI usage reveal about how people work, BioEmu2's public preview, and research aimed at improving weather forecasts weeks in advance. Watch the full video here or listen on major podcast platforms.

6:55