The Gradient: Perspectives on AI Sara Hooker: Cohere For AI, the Hardware Lottery, and DL Tradeoffs
11 snips
Jul 14, 2022 Chapters
Transcript
Episode notes
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Introduction
00:00 • 2min
Podcasts
02:02 • 4min
How Did You Become Into Machine Learning?
05:48 • 3min
Machine Learning
08:44 • 3min
The Best Ideas Don't Win
11:45 • 4min
Is There a Future for Modular Architecture?
15:18 • 2min
Benchmarks for Hardware Back Ends
17:47 • 6min
Benchmarking Hardware Lottery
23:25 • 1min
Is the Hardware Lottery Overindexing?
24:49 • 5min
How Do You Grow a Network With Unstructured Sparsity?
29:37 • 2min
Is There a Trait Off Between Sparsity and Performance?
31:53 • 3min
The Need for Evolution of Benchmarks
34:39 • 4min
Is There a Long Tail in the Model?
38:55 • 3min
How Do We Get the Best Out of Difficult Examples?
41:29 • 4min
Coher Fi Is a Lab That's Focused on Making Progress on Fundamental Research
45:11 • 2min
How Do You Hope to Start a Morl Fellowship?
47:05 • 4min
Coher for a - I'm So Excited!
50:46 • 2min