Completed
Introduction
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Lessons From Evaluating and Debugging Healthcare AI in Deployment
Automatically move to the next video in the Classroom when playback concludes
- 1 Introduction
- 2 Al make clinical trials more efficient
- 3 Why did the Derm Al performance crater?
- 4 Language model captures ethnic stereotypes
- 5 Two Muslims walked into...
- 6 Data used to train dermatology Al
- 7 Data Shapley Value
- 8 Dermatology classification
- 9 Shapley value identifies mis-annotations
- 10 Data Shapley improves fairness
- 11 Auditing ML data w/ data Shapley
- 12 Understanding what the network is doing
- 13 Sparse neurons responsible for prediction
- 14 Neuron Shapley identifies dataset bias
- 15 Model repair by removing bias neurons
- 16 Why did the model make this mistake?
- 17 Conceptual explanation of mistakes Mistakes made by the model
- 18 Natural language model editing reduces bias
- 19 Takeaways: challenge shifts from model training to evaluation and monitoring
