Clik here to view.
MatterGen: A Generative Model for Materials Design
Presented by Tian Xie at Microsoft Research Forum, June 2024 “Materials design is the cornerstone of modern technology. Many of the challenges our society is facing today are bottlenecked by finding a...
View ArticleClik here to view.
AutoGen Update: Complex Tasks and Agents
Presented by Adam Fourney at Microsoft Research Forum, June 2024 “Agents are a very, very powerful abstraction over things like task decomposition, specialization, tool use, etc. Really, you think...
View ArticleClik here to view.
MG-TSD: Advancing time series analysis with multi-granularity guided...
Author: Chang Xu Diffusion probabilistic models have the capacity to generate high-fidelity samples for generative time series forecasting. However, they also present issues of instability due to...
View ArticleClik here to view.
Direct Nash Optimization: Teaching language models to self-improve with...
Presented by Corby Rosset at Microsoft Research Forum, September 2024 “The traditional way to fine-tune an LLM for post-training … basically tells the model to emulate good behaviors, but it does not...
View ArticleClik here to view.
Project Aurora: The first large-scale foundation model of the atmosphere
Presented by Megan Stanley at Microsoft Research Forum, September 2024 “If we look at Aurora’s ability to predict pollutants such as nitrogen dioxide that are strongly related to emissions for human...
View ArticleClik here to view.
A generative model of biology for in-silico experimentation and discovery
Presented by Kevin Yang at Microsoft Research Forum, September 2024 “EvoDiff is a discrete diffusion model trained on evolutionary-scale protein sequence data. By evolutionary scale, we mean that we...
View ArticleClik here to view.
Fostering appropriate reliance on AI
Presented by Mihaela Vorvoreanu at Microsoft Research Forum, September 2024 “This is where I think it is our responsibility as people working in UX disciplines—as people researching UX and...
View ArticleClik here to view.
VALL-E 2: Enhancing the robustness and naturalness of text-to-speech models
Author: Shujie Liu In recent years, the rapid advancement of AI has continually expanded the capabilities of Text-to-Speech (TTS) technology. Ongoing optimizations and innovations in TTS have enriched...
View ArticleClik here to view.
nnScaler: Exploring a new paradigm for parallel execution in deep learning
Author: Youshan Miao Today, deep learning has permeated our daily lives. As the size of models continues to grow, training these models on massive GPU accelerators has become increasingly...
View ArticleClik here to view.
ProbTS: Unified benchmarking for time-series forecasting
Author: Machine Learning Group Time-series forecasting is crucial across various industries, including health, energy, commerce, climate, etc. Accurate forecasts over different prediction horizons are...
View ArticleClik here to view.
OmniParser for pure vision-based GUI agent
By Yadong Lu, Senior Researcher; Jianwei Yang, Principal Researcher; Yelong Shen, Principal Research Manager; Ahmed Awadallah, Partner Research Manager Recent advancements in large vision-language...
View ArticleClik here to view.
USENIX ATC 2024 best paper | How Microsoft is improving cloud AI...
As cloud AI workloads grow in complexity and scale, maintaining high system reliability has become crucial. Traditional methods of ensuring system reliability, such as using redundant components,...
View ArticleClik here to view.
Timeline: Assistive technology at Microsoft Research
Empowering every person and organization on the planet to achieve more requires giving ownership of the computing experience to the individual and leveraging technological advancements to deliver...
View ArticleClik here to view.
Magentic-One: A Generalist Multi-Agent System for Solving Complex Tasks
By Adam Fourney, Principal Researcher; Gagan Bansal, Senior Researcher; Hussein Mozannar, Senior Researcher; Victor Dibia, Principal Research Software Engineer; Saleema Amershi, Partner Research...
View ArticleClik here to view.
Experimentation in GenAI: C++ Team’s Practices for Continuous Improvement
By Sinem Akinci, Microsoft Developer Division and Cindy Chiu, Microsoft Experimentation Platform Generative AI (GenAI)[1] leverages deep learning models to identify underlying patterns and generate...
View ArticleClik here to view.
Towards a Synergy between AI and the Brain: Microsoft Research Asia StarTrack...
In a groundbreaking move, Microsoft Research Asia’s prestigious StarTrack Scholars has officially taken flight, extending a global invitation to brilliant young minds for an immersive three-month...
View ArticleClik here to view.
A generative model of biology for in-silico experimentation and discovery
Presented by Kevin Yang at Microsoft Research Forum, September 2024 “EvoDiff is a discrete diffusion model trained on evolutionary-scale protein sequence data. By evolutionary scale, we mean that we...
View ArticleClik here to view.
External Validity of Online Experiments: Can We Predict the Future?
“It is difficult to make predictions, especially about the future” – Yogi Berra (perhaps apocryphal) How well can experiments be used to predict the future? At Microsoft’s Experimentation Platform...
View ArticleNeurIPS 2024 Papers from Microsoft Research Asia
Opens in a new tab The post NeurIPS 2024 Papers from Microsoft Research Asia appeared first on Microsoft Research.
View ArticleClik here to view.
Next-Generation Systems in the AI Era: Microsoft Research Asia StarTrack...
AI is advancing at an extraordinary pace, driven by groundbreaking innovations in computer systems. As we rapidly move towards an era centered around AI computing, these systems have significantly...
View Article