Gradient Dissent: Conversations on AI

Drago Anguelov — Robustness, Safety, and Scalability at Waymo

Jul 14, 2022
Ask episode
Chapters
Transcript
Episode notes
1
Introduction
00:00 • 6min
2
What Is the Most Powerful Detection Technique Ever?
05:53 • 3min
3
Computer Vision
08:41 • 2min
4
Is There Still a Need for Image Ned?
10:28 • 2min
5
Autonomous Vehicles - What's the Inner Ilitativ Agreement?
12:31 • 5min
6
Machine Learning Models - What Are the Trends Over the Last Ten Years?
17:15 • 4min
7
How to Optimize a Tooling System?
20:48 • 3min
8
Scalability Is Always a Problem, Right?
23:57 • 2min
9
Cars and Trucks
25:56 • 2min
10
The Big Challenges of Machine Learning
27:38 • 2min
11
Do You Think Lighter Will Always Be Necessary?
29:40 • 2min
12
Is Mapping Scalable?
31:58 • 4min
13
Trucks - What's the Difference?
35:29 • 2min
14
Is There a Robustness in a Hybrid?
37:32 • 3min
15
Weymor - What's the Best Way to Engage the Community?
40:44 • 4min
16
The Challengers - A New Approach to Computer Vision
44:44 • 3min
17
Is It a Challenge to Release the Data?
47:28 • 2min
18
Machine Learning Models - What Do You Think Are Understudied?
49:02 • 3min
19
How to Train a Simulation Model to Produce Realistic Results?
51:58 • 2min
20
Is Distributional Realism Really Realistic?
53:35 • 3min
21
Do You Have More Thoughts on Finding?
56:39 • 4min
22
Is Longtail a Class Uncertainty?
01:01:01 • 3min
23
The Hardest Part of Making Autonomous Vehicles Work?
01:03:37 • 5min