95 citations · 284 across the 17 of their papers we have counts for
18 papers
De-amplifying Bias from Differential Privacy in Language Model Fine-tuning
Sanjari Srivastava, Piotr Mardziel, Zhikhun Zhang +3
Fairness and privacy are two important values machine learning (ML) practitioners often seek to operationalize in models. Fairness aims to reduce model bias for social/demographic…
Influence Patterns for Explaining Information Flow in BERT
Kaiji Lu, Zifan Wang, Piotr Mardziel +1
While attention is all you need may be proving true, we do not know why: attention-based transformer models such as BERT are superior but how information flows from input tokens to…
Reconstructing Actions To Explain Deep Reinforcement Learning
Xuan Chen, Zifan Wang, Yucai Fan +4
Feature attribution has been a foundational building block for explaining the input feature importance in supervised learning with Deep Neural Network (DNNs), but face new challeng…
Fairness Under Feature Exemptions: Counterfactual and Observational Measures
Sanghamitra Dutta, Praveen Venkatesh, Piotr Mardziel +2
With the growing use of ML in highly consequential domains, quantifying disparity with respect to protected attributes, e.g., gender, race, etc., is important. While quantifying di…
Smoothed Geometry for Robust Attribution
Zifan Wang, Haofan Wang, Shakul Ramkumar +3
Feature attributions are a popular tool for explaining the behavior of Deep Neural Networks (DNNs), but have recently been shown to be vulnerable to attacks that produce divergent…
Influence Paths for Characterizing Subject-Verb Number Agreement in LSTM Language Models
Kaiji Lu, Piotr Mardziel, Klas Leino +2
LSTM-based recurrent neural networks are the state-of-the-art for many natural language processing (NLP) tasks. Despite their performance, it is unclear whether, or how, LSTMs lear…
Interpreting Interpretations: Organizing Attribution Methods by Criteria
Zifan Wang, Piotr Mardziel, Anupam Datta +1
Motivated by distinct, though related, criteria, a growing number of attribution methods have been developed tointerprete deep learning. While each relies on the interpretability o…
Score-CAM: Score-Weighted Visual Explanations for Convolutional Neural Networks
Haofan Wang, Zifan Wang, Mengnan Du +5
Recently, increasing attention has been drawn to the internal mechanisms of convolutional neural networks, and the reason why the network makes specific decisions. In this paper, w…
Build It, Break It, Fix It: Contesting Secure Development
James Parker, Michael Hicks, Andrew Ruef +5
Typical security contests focus on breaking or mitigating the impact of buggy systems. We present the Build-it, Break-it, Fix-it (BIBIFI) contest, which aims to assess the ability…
Gender Bias in Neural Natural Language Processing
Kaiji Lu, Piotr Mardziel, Fangjing Wu +2
We examine whether neural natural language processing (NLP) systems reflect historical biases in training data. We define a general benchmark to quantify gender bias in a variety o…
Supervising Feature Influence
Shayak Sen, Piotr Mardziel, Anupam Datta +1
Causal influence measures for machine learnt classifiers shed light on the reasons behind classification, and aid in identifying influential input features and revealing their bias…
Evaluating Design Tradeoffs in Numeric Static Analysis for Java
Shiyi Wei, Piotr Mardziel, Andrew Ruef +2
Numeric static analysis for Java has a broad range of potentially useful applications, including array bounds checking and resource usage estimation. However, designing a scalable…
Latent Factor Interpretations for Collaborative Filtering
Anupam Datta, Sophia Kovaleva, Piotr Mardziel +1
Many machine learning systems utilize latent factors as internal representations for making predictions. Since these latent factors are largely uninterpreted, however, predictions…
Proxy Non-Discrimination in Data-Driven Systems
Anupam Datta, Matt Fredrikson, Gihyuk Ko +2
Machine learnt systems inherit biases against protected classes, historically disparaged groups, from training data. Usually, these biases are not explicit, they rely on subtle cor…
Use Privacy in Data-Driven Systems: Theory and Experiments with Machine Learnt Programs
Anupam Datta, Matthew Fredrikson, Gihyuk Ko +2
This paper presents an approach to formalizing and enforcing a class of use privacy properties in data-driven systems. In contrast to prior work, we focus on use restrictions on pr…
Quantifying vulnerability of secret generation using hyper-distributions (extended version)
Mário S. Alvim, Piotr Mardziel, Michael Hicks
Traditional approaches to Quantitative Information Flow (QIF) represent the adversary's prior knowledge of possible secret values as a single probability distribution. This represe…
Build It, Break It, Fix It: Contesting Secure Development
Andrew Ruef, Michael Hicks, James Parker +3
Typical security contests focus on breaking or mitigating the impact of buggy systems. We present the Build-it Break-it Fix-it BIBIFI contest which aims to assess the ability to se…
Picking vs. Guessing Secrets: A Game-Theoretic Analysis (Technical Report)
MHR Khouzani, Piotr Mardziel, Carlos Cid +1
Choosing a hard-to-guess secret is a prerequisite in many security applications. Whether it is a password for user authentication or a secret key for a cryptographic primitive, pic…