Feature Abstraction Insights
Yannic and Andrew explore the complexities of feature abstraction in machine learning, questioning the distinction between high-level and low-level features. Andrew highlights a spectrum of learned features, revealing a core of robust abstract features alongside a significant tail of non-robust, exemplar-based features. The conversation delves into the ongoing challenge of defining what constitutes a feature in various contexts, emphasizing the nuances in understanding robustness in training data.In this clip
From this podcast
Machine Learning Street Talk (MLST)
Adversarial Examples and Data Modelling - Andrew Ilyas (MIT)
Related Questions