Completed
ImageNet-9: A Fine-Grained Study Xiao Engstrom Ilyas M 2020
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
What Do Our Models Learn? - Aleksander Mądry
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 ML Research Pipeline
- 3 Concern #1: "Classic" Overfitting
- 4 Concern #2: Adaptive Overfitting
- 5 Simple Setting: Background bias
- 6 Do Backgrounds Contain Signal?
- 7 ImageNet-9: A Fine-Grained Study Xiao Engstrom Ilyas M 2020
- 8 Adversarial Backgrounds
- 9 Background-Robust Models?
- 10 Are Better Models Better?
- 11 Biases Can Be Subtle
- 12 How Are Datasets Created?
- 13 Dataset Creation in Practice
- 14 Crowdsourced Validation: A Closer Look
- 15 Prerequisite: Detailed Annotations
- 16 Restricting Relevant Labels
- 17 From Validation to Classification
- 18 Multi-Object Images
- 19 How Does This Affect Accuracy?
- 20 Which Object Do Models Predict?
- 21 Human-Based Evaluation
- 22 Dataset Replication
- 23 Case Study: ImageNet-v2
- 24 Replication Pipeline
- 25 Statistical Bias
