Greg Yang on Communicating Research, Tensor Programs, and µTransfer from The Gradient: Perspectives on AI

Greg Yang on Communicating Research, Tensor Programs, and µTransfer

1 hour 5 minutes Posted Apr 28, 2022 at 4:00 pm.

In episode 24 of The Gradient Podcast, Daniel Bashir talks to Greg Yang, senior researcher at Microsoft Research. Greg Yang’s Tensor Programs framework recently received attention for its role in the µTransfer paradigm for tuning the hyperparameters of large neural networks. Subscribe to The Gradient Podcast: Apple Podcasts | Spotify | Pocket Casts | RSSFollow The Gradient on TwitterSections:

Intro

Start in AI / Research

Fear of Math in ML

Presentation of Research

Path to MSR

Origin of Tensor Programs

Refining TP’s Presentation

The Sea of Garbage (Initializations) and the Oasis

Scaling Up Further

On Theory and Practice in Deep Learning

0:00

1:05:50

Download MP3

Share at current time

Show notes

In episode 24 of The Gradient Podcast, Daniel Bashir talks to Greg Yang, senior researcher at Microsoft Research. Greg Yang’s Tensor Programs framework recently received attention for its role in the µTransfer paradigm for tuning the hyperparameters of large neural networks. Subscribe to The Gradient Podcast: Apple Podcasts | Spotify | Pocket Casts | RSSFollow The Gradient on TwitterSections:(00:00) Intro(01:50) Start in AI / Research(05:55) Fear of Math in ML(08:00) Presentation of Research(17:35) Path to MSR(21:20) Origin of Tensor Programs(26:05) Refining TP’s Presentation(39:55) The Sea of Garbage (Initializations) and the Oasis(47:44) Scaling Up Further(55:53) On Theory and Practice in Deep Learning(01:05:28) OutroEpisode Links:* Greg’s Homepage* Greg’s Twitter* µP GitHub* Visual Intro to Gaussian Processes (Distill) Get full access to The Gradient at thegradientpub.substack.com/subscribe

← See all 147 episodes of The Gradient: Perspectives on AI