Microsoft

company

Verified

https://www.microsoft.com/en-us/research/

AI & ML interests

None defined yet.

Recent Activity

tosch22 published a model about 17 hours ago

microsoft/UserLM-2.0

abanerjeeml updated a dataset about 22 hours ago

microsoft/SynthDocQA

xiaoyizhang-microsoft updated a model 1 day ago

microsoft/Mage-VL-Cowork

View all activity

Papers

ThinkBit: Search Codewords, Recalibrate Scales, Select by Reasoning

Breaking Diffusion: Parallel Reasoning with Draft Refinement

View all Papers

Articles

Turning Database Grading into RL Rewards: Post-Training Agents with ThinkingBox

The Agent Said It Was Done. The Database Disagreed.

Hugging Face Models on Foundry Managed Compute

Differential Transformer V2

Introducing OptiMind, a research model designed for optimization

View all articles

microsoft 's papers 91

Submitted by

Senqiao Yang

Mage-VL: An Efficient Codec-Native Streaming Multimodal Foundation Model

microsoft

Submitted by

Baolin Peng

OpenForgeRL: Train Harness-native Agents in Any Environment

microsoft

Submitted by

Xinjie

Mage-Flow: An Efficient Native-Resolution Foundation Model for Image Generation and Editing

microsoft

2

DocAtlas: Long-Document Understanding as Mutable-State Interaction

microsoft

XL-DocBench: Benchmarking Evidence-Grounded Extra-Long Document Understanding

microsoft

Submitted by

yangyu huang

ResearchStudio-Reel: Automate the Last Mile of Research from Paper to Poster, Video, and Blog

microsoft

Submitted by

Qihao Zhao

ResearchStudio-Idea: An Evidence-Grounded Research-Ideation Skill Suite from ML Conference Outcomes

microsoft

Submitted by

Amirhossein Abaskohi

SeKV: Resolution-Adaptive KV Cache with Hierarchical Semantic Memory for Long-Context LLM Inference

microsoft

Submitted by

qianchu liu

HealthAgentBench: A Unified Benchmark Suite of Realistic Agentic Healthcare Environments for Challenging Frontier AI Agents

microsoft

Submitted by

Yanuo Ma

Building to the Test: Coding Agents Deliver What You Check, Not What You Requested

microsoft

Temporal Preference Optimization for Unsupervised Retrieval

microsoft

Submitted by

Shaoqiu Zhang

FastContext: Training Efficient Repository Explorer for Coding Agents

microsoft

6