Edward Raff

Robust Design of Deep Neural Networks Against Adversarial Attacks Based on Lyapunov Theory

2020

Deep neural networks (DNNs) are vulnerable to subtle adversarial perturbations applied to the input. These adversarial perturbations, though imperceptible, can easily mislead the DNN. In this work, we take a control theoretic approach to the …

Learning with Holographic Reduced Representations

2021 · arXiv (Cornell University)

Holographic Reduced Representations (HRR) are a method for performing symbolic AI on top of real-valued vectors by associating each vector with an abstract concept, and providing mathematical operations to manipulate vectors as if they were …

Deploying Convolutional Networks on Untrusted Platforms Using 2D Holographic Reduced Representations

2022 · arXiv (Cornell University)

Due to the computational cost of running inference for a neural network, the need to deploy the inferential steps on a third party's compute environment or hardware is common. If the third party is not …

You Don't Need Robust Machine Learning to Manage Adversarial Attack Risks

2023 · arXiv (Cornell University)

The robustness of modern machine learning (ML) models has become an increasing concern within the community. The ability to subvert a model into making errant predictions using seemingly inconsequential changes to input is startling, as …

Probing the Transition to Dataset-Level Privacy in ML Models Using an Output-Specific and Data-Resolved Privacy Profile

2023 · arXiv (Cornell University)

Differential privacy (DP) is the prevailing technique for protecting user data in machine learning models. However, deficits to this framework include a lack of clarity for selecting the privacy budget $ε$ and a lack of …

Measuring Equality in Machine Learning Security Defenses: A Case Study in Speech Recognition

2023

Over the past decade, the machine learning security community has developed a myriad of defenses for evasion attacks. An understudied question in that community is: for whom do these defenses defend? This work considers common …

Neural Normalized Compression Distance and the Disconnect Between Compression and Classification

2024 · arXiv (Cornell University)

It is generally well understood that predictive classification and compression are intrinsically related concepts in information theory. Indeed, many deep learning methods are explained as learning a kind of compression, and that better compression leads …

Crosslingual Generalization through Multitask Finetuning

2023

Niklas Muennighoff, Thomas Wang, Lintang Sutawika, Adam Roberts, Stella Biderman, Teven Le Scao, M Saiful Bari, Sheng Shen, Zheng Xin Yong, Hailey Schoelkopf, Xiangru Tang, Dragomir Radev, Alham Fikri Aji, Khalid Almubarak, Samuel Albanie, Zaid …