Generally Intelligent

Episode 15: Martín Arjovsky, INRIA, on benchmarks for robustness and geometric information theory

Oct 15, 2021

01:26:13

Creator website

Highlights

AI Chapters

Episode notes

Then You Throw Away Data Points

04:01

What's the Half That's Missing?

03:47

The Best Researches That You've Worked With?

01:58

Introduction

2min

How Did You Become Interested in Machine Learning?

4min

How Machine Learning Can Reduce the Compression of Bidio Compressors?

2min

The Last Chain in the Story, Vasically

3min

Is There a Deadline for a Clear in Six Weeks?

2min

How to Make a Machine Learning Machine

2min

The Invaliant Causal Prediction Paper

2min

Is There a Gray in the Sand?

4min

The Distribution of Locations Is Not the Same

2min

10.

Using a Dataset to Find Markers for Canser in Space

4min

11.

We Need More Benchmarks on Robastness

4min

12.

Is There a Benchmark for Out of Solution Gennization?

2min

13.

How to Get a 100 Percent Curacy in a Linear Problem?

2min

14.

Then You Throw Away Data Points

4min

15.

What Makes a Bench Mark Take Off?

2min

16.

How to Carry a Good Benchmark?

5min

17.

Is the Entropy of a Negative Infinite?

3min

18.

What's the Half That's Missing?

4min

19.

Nour Networks Are Compromising With Entropy

4min

20.

High Blood Pressure - Cause and Effect

4min

21.

Papers That Go From Intuition to Solid Toy Tasks

2min

22.

The Best Papers From Shorty Ofr So Many

3min

23.

The Uncertainty in Prediction

2min

24.

I Think Scale Is Not the Answer

4min

25.

Is There a Difference Between Ganis and Unsupervised Learning?

2min

26.

How to Make a Product Work for You

2min

27.

How to Learn the Rail Without a Value Function?

2min

28.

The Best Researches That You've Worked With?

2min

29.

Working on a Porn, Is It Important?

4min

30.

Exploration in Infant Psychology

2min

31.

The Importance of Uncertainty Estimation and Nanomalyetaction

2min

Martín Arjovsky did his Ph.D. at NYU with Leon Bottou. Some of his well-known works include the Wasserstein GAN and a paradigm called Invariant Risk Minimization. In this episode, we discuss out-of-distribution generalization, geometric information theory, and the importance of good benchmarks.

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.