Publications (9)
What do Vision Transformers Learn? A Visual Exploration
Amin Ghiasi, Hamid Kazemi, Eitan Borgnia +5
Vision transformers (ViTs) are quickly becoming the de-facto architecture for computer vision, yet we understand very little about why they work and what they learn. While existing…
A Deep Dive into Dataset Imbalance and Bias in Face Identification
Valeriia Cherepanova, Steven Reich, Samuel Dooley +3
As the deployment of automated face recognition (FR) systems proliferates, bias in these systems is not just an academic question, but a matter of public concern. Media portrayals…
Class Group Relations in a Function Field Analogue of
Steven Reich
For an odd prime and polynomial , we consider the extension of defined by adjoining a root of . Such a field is a function field ana…
WITCHcraft: Efficient PGD attacks with random step size
Ping-Yeh Chiang, Jonas Geiping, Micah Goldblum +4
State-of-the-art adversarial attacks on neural networks use expensive iterative methods and numerous random restarts from different initial points. Iterative FGSM-based methods wit…
The impact of short term synaptic depression and stochastic vesicle dynamics on neural variability
Steven Reich, Robert Rosenbaum
Neural variability plays a central role in neural coding and neuronal network dynamics. Unreliability of synaptic transmission is a major source of neural variability: synaptic neu…
Unraveling Meta-Learning: Understanding Feature Representations for Few-Shot Tasks
Micah Goldblum, Steven Reich, Liam Fowl +3
Meta-learning algorithms produce feature extractors which achieve state-of-the-art performance on few-shot classification. While the literature is rich with meta-learning methods,…
Plug-In Inversion: Model-Agnostic Inversion for Vision with Data Augmentations
Amin Ghiasi, Hamid Kazemi, Steven Reich +3
Existing techniques for model inversion typically rely on hard-to-tune regularizers, such as total variation or feature regularization, which must be individually calibrated for ea…
Ensemble Distillation for Structured Prediction: Calibrated, Accurate, Fast-Choose Three
Steven Reich, David Mueller, Nicholas Andrews
Modern neural networks do not always produce well-calibrated predictions, even when trained with a proper scoring function such as cross-entropy. In classification settings, simple…
Decepticons: Corrupted Transformers Breach Privacy in Federated Learning for Language Models
Liam Fowl, Jonas Geiping, Steven Reich +4
A central tenet of Federated learning (FL), which trains models without centralizing user data, is privacy. However, previous work has shown that the gradient updates used in FL ca…