Description
The personal website of Yang Song, a prominent researcher in the field of artificial intelligence. Yang Song currently serves as the Research Principal at Meta Superintelligence Labs (MSL), where he plays a key role in shaping the lab's research direction alongside Chief Scientist Shengjia Zhao. Prior to his role at Meta, he led the strategic explorations team at OpenAI, contributing to the advancement of AI technologies.
Read More
Yang Song earned his Ph.D. in Computer Science from Stanford University, under the advisement of Stefano Ermon. During his doctoral studies, he developed foundational concepts and techniques in score-based diffusion models, which have significantly impacted the field of generative AI. His work in this area has been recognized in various publications, including a blog post, a Quanta Magazine article, and interviews. He also completed research internships at industry leaders such as Google Brain, Uber ATG, and Microsoft Research.
His academic journey began at Tsinghua University, where he obtained his Bachelor's degree in Mathematics and Physics. There, he collaborated with notable researchers including Jun Zhu, Raquel Urtasun, and Richard Zemel. His research interests lie in developing novel generative models, with a particular emphasis on diffusion models and consistency models. These models are capable of generating high-quality data, such as images, with remarkable speed and efficiency. His work has led to state-of-the-art results in image generation, inverse problem solving in medical imaging, and accelerating feedforward computations in neural networks.
Yang Song's publications showcase his contributions to major AI conferences, including ICLR, ICML, and NeurIPS. His research has explored areas such as simplifying, stabilizing, and scaling continuous-time consistency models, improving techniques for training consistency models, and developing score-based generative modeling through stochastic differential equations. He has also investigated applications of these models in solving inverse problems in medical imaging and accelerating feedforward computation via parallel nonlinear equation solving. His work on unrestricted adversarial examples and using generative models for defense against adversarial attacks further highlights the breadth of his research contributions.
- CategoryOther AI Tools
- DeploymentCloud
- PlatformsWeb
Score SDE's Core Features
Research Principal at Meta Superintelligence Labs
Former leader at OpenAI's strategic explorations team
Pioneering work in score-based diffusion models
Developed foundational concepts in continuous-time consistency models
Expertise in generative AI and probabilistic modeling
Contributions to fast sampling techniques for generative models
Applications in image generation and medical imaging reconstruction
Research on accelerating neural network computations
Authored numerous influential publications in top AI conferences
Ph.D. in Computer Science from Stanford University
Getting Started with Score SDE
Access Research: Explore publications and research papers on generative models.
Understand Models: Learn about diffusion models and consistency models.
Explore Code: View code repositories for research projects.
Contact Researcher: Reach out for collaboration or inquiries.
Score SDE's Use Cases
- Generative AI Research
- Image Generation
- Medical Imaging
- Model Optimization
- AI Strategy
