The Gradient: Perspectives on AI

Been Kim: Interpretable Machine Learning

Aug 18, 2022

01:11:32

Creator website

Highlights

AI Chapters

Episode notes

Interprotability Methods - What's Your Biggest Worry?

02:30

How Can We Use a Eye to Help People Be More Creative?

03:29

The Difficulty of Causal Inference

03:45

How Different Architectures Influence Different Concepts?

03:14

Introduction

3min

AlexNet

2min

Alot's Net: A PhD in Interpretability

2min

Inherently Interpretable Models Are One Way Communication

2min

Human Experiments - A Toy Showcase of ML Interpretability

2min

Machine Learning - The Golden Standard

2min

Towards a Rigid Science of Interpretability

2min

The Issues That I've Seen at the Time

2min

The Interpretability Field Is Taking Similar Path

4min

10.

Is There a Confirmation Bias in Machine Learning?

2min

11.

Sanity Check for a Sanity and C Map Paper

3min

12.

Machine Learning - What's Your Worst Fear?

2min

13.

Is There a Language Between Humans and Machines?

2min

14.

The Relationship Between Humans and Machines

3min

15.

Cross Training and Mental Model

2min

16.

Is There a Shift in Interpretability?

2min

17.

Adaptation and Deficiency in the Medical Field

2min

18.

TCAV Methods - Concept Based Explanation

3min

19.

The Power of T-Cab and the Directional Derivative

2min

20.

T-Cab - Concept Based Explanation

2min

21.

How Can We Use AI to Help People Be More Creative?

2min

22.

What's Your Concept of Calmness?

2min

23.

T-Cav - Can We Discover Concepts?

3min

24.

How to Improve the Design of a Neural Network?

2min

25.

How Different Architectures Influence Different Concepts?

3min

26.

Alpha Zero - Is There Even an Overlap Between Alpha 0 and How Humans Play Chess?

2min

27.

Alpha Zero Picks Up Human Concepts Like Material Unbalance

2min

28.

How Do Human Chess Players Learn Differently?

2min

29.

How Can the Fuzzero Surprise Us in a Way That's Productive?

3min

30.

The Importance of Having Experts in the Room in Machine Learning

3min

31.

Is Your Work Impacting the World?

2min

32.

Don't Change Who You Are

3min

In episode 38 of The Gradient Podcast, Daniel Bashir speaks to Been Kim.

Been is a staff research scientist at Google Brain focused on interpretability–helping humans communicate with complex machine learning models by not only building tools but also studying how humans interact with these systems. She has served with a number of conferences including ICLR, NeurIPS, ICML, and AISTATS. She gave the keynotes at ICLR 2022, ECML 2020, and the G20 meeting in Argentina in 2018. Her work TCAV received the UNESCO Netexplo award, was featured at Google I/O 2019 and in Brian Christian’s book The Alignment Problem.

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

Outline:

(00:00) Intro(02:20) Path to AI/interpretability(06:10) The Progression of Been’s thinking / PhD thesis(11:30) Towards a Rigorous Science of Interpretable Machine Learning(24:52) Interpretability and Software Testing(27:00) Been’s ICLR Keynote and Human-Machine “Language”(37:30) TCAV(43:30) Mood Board Search and CAV Camera(48:00) TCAV’s Limitations and Follow-up Work(56:00) Acquisition of Chess Knowledge in AlphaZero(1:07:00) Daniel spends a very long time asking “what does it mean to you to be a researcher?”(1:09:00) The everyday drudgery, more lessons from Been(1:11:32) Outro

Links:

* Been’s website

* CAVcamera app

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