Dawn Song

New Directions in Social Authentication

2015

Author(s): Sakshi Jain, Juan Lang, Neil Zhenqiang Gong, Dawn Song, Sreya Basuroy, Prateek Mittal Download: Paper (PDF) Date: 7 Feb 2015 Document Type: Briefing Papers Additional Documents: Slides Associated Event: NDSS Symposium 2015 Abstract: Web …

Robust Physical-World Attacks on Machine Learning Models.

2017 · arXiv (Cornell University)

Deep neural network-based classifiers are known to be vulnerable to adversarial examples that can fool them into misclassifying their input through the addition of small-magnitude perturbations. However, recent studies have demonstrated that such adversarial examples …

Generating Adversarial Examples with Adversarial Networks

2018 · arXiv (Cornell University)

Deep neural networks (DNNs) have been found to be vulnerable to adversarial examples resulting from adding small-magnitude perturbations to inputs. Such adversarial examples can mislead DNNs to produce adversary-selected results. Different attack strategies have been …

Adversarial Texts with Gradient Methods

2018 · arXiv (Cornell University)

Adversarial samples for images have been extensively studied in the literature. Among many of the attacking methods, gradient-based methods are both effective and easy to compute. In this work, we propose a framework to adapt …

Learning what to learn in a neural program

2018

Learning programs with neural networks is a challenging task, addressed by a long line of existing work. It is difficult to learn neural networks which will generalize to problem instances that are much larger than …

Characterizing Adversarial Subspaces Using Local Intrinsic Dimensionality

2018 · Own your potential (DEAKIN)

© Learning Representations, ICLR 2018 - Conference Track Proceedings.All right reserved. Deep Neural Networks (DNNs) have recently been shown to be vulnerable against adversarial examples, which are carefully crafted instances that can mislead DNNs to …

F^2ed-Learning: Good Fences Make Good Neighbors

2021 · arXiv (Cornell University)

In this paper, we present F^2ed-Learning, the first federated learning protocol simultaneously defending against both semi-honest server and Byzantine malicious clients. Using a robust mean estimator called FilterL2, F^2ed-Learning is the first FL protocol with …

What Would Jiminy Cricket Do? Towards Agents That Behave Morally

2021 · arXiv (Cornell University)

When making everyday decisions, people are guided by their conscience, an internal sense of right and wrong. By contrast, artificial agents are currently not endowed with a moral sense. As a consequence, they may learn …

Grounded Graph Decoding Improves Compositional Generalization in Question Answering

2021 · arXiv (Cornell University)

Question answering models struggle to generalize to novel compositions of training patterns, such to longer sequences or more complex test structures. Current end-to-end models learn a flat input embedding which can lose input syntax context. …

COSMIC: Generalized Refusal Direction Identification in LLM Activations

2025 · arXiv (Cornell University)

Large Language Models (LLMs) encode behaviors such as refusal within their activation space, yet identifying these behaviors remains a significant challenge. Existing methods often rely on predefined refusal templates detectable in output tokens or require …

SQLNet: Generating Structured Queries From Natural Language Without Reinforcement Learning

2017 · arXiv (Cornell University)

Synthesizing SQL queries from natural language is a long-standing open problem and has been attracting considerable interest recently. Toward solving the problem, the de facto approach is to employ a sequence-to-sequence-style model. Such an approach …

Measuring Massive Multitask Language Understanding

2020 · arXiv (Cornell University)

We propose a new test to measure a text model's multitask accuracy. The test covers 57 tasks including elementary mathematics, US history, computer science, law, and more. To attain high accuracy on this test, models …

Measuring Massive Multitask Language Understanding

2021 · International Conference on Learning Representations

We propose a new test to measure a text model's multitask accuracy. The test covers 57 tasks including elementary mathematics, US history, computer science, law, and more. To attain high accuracy on this test, models …