Sara Hooker | Research at Google AI | Interpretability and Model Compression | CTDS.Show #99

Published: Sept. 10, 2020, 3:59 p.m.

Video Version: https://youtu.be/O7JwKAv99_M

\n

Subscribe here to the newsletter: https://tinyletter.com/sanyambhutani

\n

In this episode, Sanyam Bhutani interviews Research Scholar at Google brain: Sara Hooker.

\n

Sara has an amazing background and a "non-traditional" path into research. They talk all about her journey into research into machine learning. Sara is still very much connected to her roots, she's from Africa, and she shares her amazing perspective on how different problems, different constraints can contribute to the field, research, or otherwise in interesting ways.

\n

Towards the end they also briefly dive into her recent works.

\n

Links:

\n

Underrated ML: https://www.underratedml.com

\n

Selective Brain Damage: Measuring the Disparate Impact of Model Pruning: https://arxiv.org/pdf/1911.05248

\n

Follow:

\n

Sara Hooker:

\n

https://twitter.com/sarahookr

\n

https://www.sarahooker.me

\n

Sanyam Bhutani:

\n

https://twitter.com/bhutanisanyam1

\n

Blog: sanyambhutani.com

\n

About:

\n

https://sanyambhutani.com/tag/chaitimedatascience/

\n

A show for Interviews with Practitioners, Kagglers & Researchers and all things Data Science hosted by Sanyam Bhutani.

\n

You can expect weekly episodes every available as Video, Podcast, and blogposts.

\n

If you'd like to support the podcast: https://www.patreon.com/chaitimedatascience

\n

Intro track:

\n

Flow by LiQWYD https://soundcloud.com/liqwyd