Lightning talks: Training and inference efficiency

Channel:
Microsoft Research
Subscribers:
343,000
Published on ● Video Link: https://www.youtube.com/watch?v=hHyjmW-EVjA


Duration: 50:20
208 views
1

To bring AI to more people, models need to be cheaper to train and run, in terms of both computational and human resources. Increasing efficiency across various parts of the training and inference pipeline includes optimizing existing large models and creating new architectures and training paradigms.

#MSFTResearchSummit

See related sessions in this track: https://www.microsoft.com/en-us/research/video/keynote-with-guests-toward-ai-that-empowers-more-people-more-of-the-time/

Learn more about the 2022 Microsoft Research Summit: https://www.microsoft.com/en-us/research/event/microsoft-research-summit-2022/



Other Videos By Microsoft Research


2022-10-27Panel discussion: Towards climate-smart cities: IoT networks for air pollution sensing
2022-10-27Lightning talks: Sustainably nourishing the world
2022-10-27Panel discussion and research talk: Computational advances in climate risk assessment
2022-10-27Keynote with guests: Accelerating precision health through scientific innovation and discovery
2022-10-27Panel discussion: Ambient clinical intelligence the next frontier of AI
2022-10-27Research talk: Differentially private fine-tuning of large language models
2022-10-27Lightning talks: Responsible AI: The challenge of big models
2022-10-27Research talk: Cloud Intelligence/AIOps – Infusing AI into cloud computing
2022-10-27Panel discussion: Emerging computing technologies in academia and industry
2022-10-27Keynote with guests: Toward AI that empowers more people more of the time
2022-10-27Lightning talks: Training and inference efficiency
2022-10-27Lightning talks: Skills acquisition and new capabilities
2022-10-27Lightning talks: Aligning models with human intent
2022-10-27Fireside chat: What’s next in large-scale AI
2022-10-27Plenary: Emerging foundations for planet-scale computing [ASL version]
2022-10-27Panel discussion: From AI research to application in AAA-games
2022-10-27Panel discussion: Is the future of the metaverse work or play?
2022-10-27Lightning talks: Empowering mathematicians with technology
2022-10-27Research talk: Numerical weak AI and symbolic strong AI
2022-10-27Plenary: The fifth paradigm of scientific discovery [ASL version]
2022-10-27Responsible, Equitable, and Ethical AI panel discussion


Tags:
Ahmed Awadallah
Artificial intelligence
Large-Scale AI
Lightning talks: Training and inference efficiency | T602
Programming languages & software engineering
Song Han
Subho Mukherjee
Yuxiong He
microsoft research summit 2022
ms research summit
msft resarch summit 2022
msft summit
msft summit 2022
research summit
summit 2022