Black Boxes Are Not Required
Deep neural networks are undeniably effective. They rely on such a high number of parameters, that they are appropriately described as "black boxes". While black boxes lack desirably properties like interpretability and explainability, in some cases, their accuracy makes them incredibly useful. But does achiving "usefulness" require a black box? Can we be sure an equally valid but simpler solution does not exist? Cynthia Rudin helps us answer that question. We discuss her recent paper with co-author Joanna Radin titled (spoiler warning)… Why Are We Using Black Box Models in AI When We Don't Need To? A Lesson From An Explainable AI Competition
5 Jun 202032min
Robustness to Unforeseen Adversarial Attacks
Daniel Kang joins us to discuss the paper Testing Robustness Against Unforeseen Adversaries.
30 Mai 202021min
Estimating the Size of Language Acquisition
Frank Mollica joins us to discuss the paper Humans store about 1.5 megabytes of information during language acquisition
22 Mai 202025min
Interpretable AI in Healthcare
Jayaraman Thiagarajan joins us to discuss the recent paper Calibrating Healthcare AI: Towards Reliable and Interpretable Deep Predictive Models.
15 Mai 202035min
Understanding Neural Networks
What does it mean to understand a neural network? That's the question posted on this arXiv paper. Kyle speaks with Tim Lillicrap about this and several other big questions.
8 Mai 202034min
Self-Explaining AI
Dan Elton joins us to discuss self-explaining AI. What could be better than an interpretable model? How about a model wich explains itself in a conversational way, engaging in a back and forth with the user. We discuss the paper Self-explaining AI as an alternative to interpretable AI which presents a framework for self-explainging AI.
2 Mai 202032min
Self Driving Cars and Pedestrians
We are joined by Arash Kalatian to discuss Decoding pedestrian and automated vehicle interactions using immersive virtual reality and interpretable deep learning.
18 Apr 202030min
