Hattie Zhou: Lottery Tickets and Algorithmic Reasoning in LLMs

Hattie Zhou: Lottery Tickets and Algorithmic Reasoning in LLMs

Author: Daniel Bashir February 16, 2023 Duration: 1:42:59

In episode 60 of The Gradient Podcast, Daniel Bashir speaks to Hattie Zhou.

Hattie is a PhD student at the Université de Montréal and Mila. Her research focuses on understanding how and why neural networks work, based on the belief that the performance of modern neural networks exceeds our understanding and that building more capable and trustworthy models requires bridging this gap. Prior to Mila, she spent time as a data scientist at Uber and did research with Uber AI Labs.

Have suggestions for future podcast guests (or other feedback)? Let us know here!

Subscribe to The Gradient Podcast:  Apple Podcasts  | Spotify | Pocket Casts | RSSFollow The Gradient on Twitter

Outline:

* (00:00) Intro

* (01:55) Hattie’s Origin Story, Uber AI Labs, empirical theory and other sorts of research

* (10:00) Intro to the Lottery Ticket Hypothesis & Deconstructing Lottery Tickets

* (14:30) Lottery tickets as lucky initialization

* (17:00) Types of masking and the “masking is training” claim

* (24:00) Type-0 masks and weight evolution over long training trajectories

* (27:00) Can you identify good masks or training trajectories a priori?

* (29:00) The role of signs in neural net initialization

* (35:27) The Supermask

* (41:00) Masks to probe pretrained models and model steerability

* (47:40) Fortuitous Forgetting in Connectionist Networks

* (54:00) Relationships to other work (double descent, grokking, etc.)

* (1:01:00) The iterative training process in fortuitous forgetting, scale and value of exploring alternatives

* (1:03:35) In-Context Learning and Teaching Algorithmic Reasoning

* (1:09:00) Learning + algorithmic reasoning, prompting strategy

* (1:13:50) What’s happening with in-context learning?

* (1:14:00) Induction heads

* (1:17:00) ICL and gradient descent

* (1:22:00) Algorithmic prompting vs discovery

* (1:24:45) Future directions for algorithmic prompting

* (1:26:30) Interesting work from NeurIPS 2022

* (1:28:20) Hattie’s perspective on scientific questions people pay attention to, underrated problems

* (1:34:30) Hattie’s perspective on ML publishing culture

* (1:42:12) Outro

Links:

* Hattie’s homepage and Twitter

* Papers

* Deconstructing Lottery Tickets: Zeros, signs, and the Supermask

* Fortuitous Forgetting in Connectionist Networks

* Teaching Algorithmic Reasoning via In-context Learning



Get full access to The Gradient at thegradientpub.substack.com/subscribe

Hosted by Daniel Bashir, The Gradient: Perspectives on AI moves beyond surface-level headlines to explore the intricate machinery and human ideas shaping artificial intelligence. Each episode is built on a foundation of deep research, leading to conversations that are both technically substantive and broadly accessible. You'll hear from researchers, engineers, and philosophers who are actively building and critiquing our technological future, discussing not just how AI systems work, but the larger implications of their integration into society. This isn't about speculative hype; it's a grounded examination of real progress, persistent challenges, and ethical considerations from those on the front lines. The discussions peel back layers on topics like model architecture, policy, and the fundamental science behind the algorithms becoming part of our daily lives. For anyone curious about the substance behind the buzz-whether you have a technical background or are simply keen to understand a defining technology of our age-this podcast offers a crucial and thoughtful resource. Tune in for a consistently detailed and nuanced take that treats artificial intelligence with the complexity it deserves.
Author: Language: English Episodes: 100

The Gradient: Perspectives on AI
Podcast Episodes
Judy Fan: Reverse Engineering the Human Cognitive Toolkit [not-audio_url] [/not-audio_url]

Duration: 1:32:39
Episode 136I spoke with Judy Fan about:* Our use of physical artifacts for sensemaking* Why cognitive tools can be a double-edged sword* Her approach to scientific inquiry and how that approach has developedEnjoy—and let…
L.M. Sacasas: The Questions Concerning Technology [not-audio_url] [/not-audio_url]

Duration: 1:47:20
Episode 135I spoke with L. M. Sacasas about:* His writing and intellectual influences* The value of asking hard questions about technology and our relationship to it* What happens when we decide to outsource skills and c…
Pete Wolfendale: The Revenge of Reason [not-audio_url] [/not-audio_url]

Duration: 2:52:57
Episode 134I spoke with Pete Wolfendale about:* The flaws in longtermist thinking* Selections from his new book, The Revenge of Reason* Metaphysics* What philosophy has to say about reason and AIEnjoy—and let me know wha…
Peter Lee: Computing Theory and Practice, and GPT-4's Impact [not-audio_url] [/not-audio_url]

Duration: 1:01:48
Episode 133I spoke with Peter Lee about:* His early work on compiler generation, metacircularity, and type theory* Paradoxical problems* GPT-4s impact, Microsoft’s “Sparks of AGI” paper, and responses and criticismEnjoy—…
Manuel & Lenore Blum: The Conscious Turing Machine [not-audio_url] [/not-audio_url]

Duration: 2:23:04
Episode 132I spoke with Manuel and Lenore Blum about:* Their early influences and mentors* The Conscious Turing Machine and what theoretical computer science can tell us about consciousnessEnjoy—and let me know what you…
Kevin Dorst: Against Irrationalist Narratives [not-audio_url] [/not-audio_url]

Duration: 2:15:21
Episode 131I spoke with Professor Kevin Dorst about:* Subjective Bayesianism and epistemology foundations* What happens when you’re uncertain about your evidence* Why it’s rational for people to polarize on political mat…
David Pfau: Manifold Factorization and AI for Science [not-audio_url] [/not-audio_url]

Duration: 2:00:52
Episode 130I spoke with David Pfau about:* Spectral learning and ML* Learning to disentangle manifolds and (projective) representation theory* Deep learning for computational quantum mechanics* Picking and pursuing resea…
Sergiy Nesterenko: Automating Circuit Board Design [not-audio_url] [/not-audio_url]

Duration: 1:03:35
Episode 128I spoke with Sergiy Nesterenko about:* Developing an automated system for designing PCBs* Difficulties in human and automated PCB design* Building a startup at the intersection of different areas of expertiseB…