Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox
Picture for Dawn Song

Dawn Song

University of California, Berkeley

BACKDOORL: Backdoor Attack against Competitive Reinforcement Learning


May 02, 2021
Lun Wang, Zaynah Javed, Xian Wu, Wenbo Guo, Xinyu Xing, Dawn Song

Figure 1 for BACKDOORL: Backdoor Attack against Competitive Reinforcement Learning
Figure 2 for BACKDOORL: Backdoor Attack against Competitive Reinforcement Learning
Figure 3 for BACKDOORL: Backdoor Attack against Competitive Reinforcement Learning
Figure 4 for BACKDOORL: Backdoor Attack against Competitive Reinforcement Learning

  Access Paper or Ask Questions

Model-Contrastive Federated Learning


Mar 30, 2021
Qinbin Li, Bingsheng He, Dawn Song

Figure 1 for Model-Contrastive Federated Learning
Figure 2 for Model-Contrastive Federated Learning
Figure 3 for Model-Contrastive Federated Learning
Figure 4 for Model-Contrastive Federated Learning
* Accepted by CVPR 2021 
  Access Paper or Ask Questions

Measuring Mathematical Problem Solving With the MATH Dataset


Mar 05, 2021
Dan Hendrycks, Collin Burns, Saurav Kadavath, Akul Arora, Steven Basart, Eric Tang, Dawn Song, Jacob Steinhardt

Figure 1 for Measuring Mathematical Problem Solving With the MATH Dataset
Figure 2 for Measuring Mathematical Problem Solving With the MATH Dataset
Figure 3 for Measuring Mathematical Problem Solving With the MATH Dataset
Figure 4 for Measuring Mathematical Problem Solving With the MATH Dataset
* Code and the MATH dataset is available at https://github.com/hendrycks/math/ 
  Access Paper or Ask Questions

DPlis: Boosting Utility of Differentially Private Deep Learning via Randomized Smoothing


Mar 02, 2021
Wenxiao Wang, Tianhao Wang, Lun Wang, Nanqing Luo, Pan Zhou, Dawn Song, Ruoxi Jia

Figure 1 for DPlis: Boosting Utility of Differentially Private Deep Learning via Randomized Smoothing
Figure 2 for DPlis: Boosting Utility of Differentially Private Deep Learning via Randomized Smoothing
Figure 3 for DPlis: Boosting Utility of Differentially Private Deep Learning via Randomized Smoothing
Figure 4 for DPlis: Boosting Utility of Differentially Private Deep Learning via Randomized Smoothing
* 20 pages, 7 figures 
  Access Paper or Ask Questions

TeraPipe: Token-Level Pipeline Parallelism for Training Large-Scale Language Models


Feb 16, 2021
Zhuohan Li, Siyuan Zhuang, Shiyuan Guo, Danyang Zhuo, Hao Zhang, Dawn Song, Ion Stoica

Figure 1 for TeraPipe: Token-Level Pipeline Parallelism for Training Large-Scale Language Models
Figure 2 for TeraPipe: Token-Level Pipeline Parallelism for Training Large-Scale Language Models
Figure 3 for TeraPipe: Token-Level Pipeline Parallelism for Training Large-Scale Language Models
Figure 4 for TeraPipe: Token-Level Pipeline Parallelism for Training Large-Scale Language Models

  Access Paper or Ask Questions

A System for Automated Open-Source Threat Intelligence Gathering and Management


Jan 19, 2021
Peng Gao, Xiaoyuan Liu, Edward Choi, Bhavna Soman, Chinmaya Mishra, Kate Farris, Dawn Song

Figure 1 for A System for Automated Open-Source Threat Intelligence Gathering and Management
Figure 2 for A System for Automated Open-Source Threat Intelligence Gathering and Management

  Access Paper or Ask Questions

A System for Efficiently Hunting for Cyber Threats in Computer Systems Using Threat Intelligence


Jan 17, 2021
Peng Gao, Fei Shao, Xiaoyuan Liu, Xusheng Xiao, Haoyuan Liu, Zheng Qin, Fengyuan Xu, Prateek Mittal, Sanjeev R. Kulkarni, Dawn Song

Figure 1 for A System for Efficiently Hunting for Cyber Threats in Computer Systems Using Threat Intelligence
Figure 2 for A System for Efficiently Hunting for Cyber Threats in Computer Systems Using Threat Intelligence
Figure 3 for A System for Efficiently Hunting for Cyber Threats in Computer Systems Using Threat Intelligence
* Accepted paper at ICDE 2021 demonstrations track. arXiv admin note: substantial text overlap with arXiv:2010.13637 
  Access Paper or Ask Questions

Dataset Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses


Dec 30, 2020
Micah Goldblum, Dimitris Tsipras, Chulin Xie, Xinyun Chen, Avi Schwarzschild, Dawn Song, Aleksander Madry, Bo Li, Tom Goldstein

Figure 1 for Dataset Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses
Figure 2 for Dataset Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses
Figure 3 for Dataset Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses
Figure 4 for Dataset Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses

  Access Paper or Ask Questions

Data Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses


Dec 18, 2020
Micah Goldblum, Dimitris Tsipras, Chulin Xie, Xinyun Chen, Avi Schwarzschild, Dawn Song, Aleksander Madry, Bo Li, Tom Goldstein

Figure 1 for Data Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses
Figure 2 for Data Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses
Figure 3 for Data Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses
Figure 4 for Data Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses

  Access Paper or Ask Questions

Extracting Training Data from Large Language Models


Dec 14, 2020
Nicholas Carlini, Florian Tramer, Eric Wallace, Matthew Jagielski, Ariel Herbert-Voss, Katherine Lee, Adam Roberts, Tom Brown, Dawn Song, Ulfar Erlingsson, Alina Oprea, Colin Raffel

Figure 1 for Extracting Training Data from Large Language Models
Figure 2 for Extracting Training Data from Large Language Models
Figure 3 for Extracting Training Data from Large Language Models
Figure 4 for Extracting Training Data from Large Language Models

  Access Paper or Ask Questions

Towards Defending Multiple Adversarial Perturbations via Gated Batch Normalization


Dec 03, 2020
Aishan Liu, Shiyu Tang, Xianglong Liu, Xinyun Chen, Lei Huang, Zhuozhuo Tu, Dawn Song, Dacheng Tao

Figure 1 for Towards Defending Multiple Adversarial Perturbations via Gated Batch Normalization
Figure 2 for Towards Defending Multiple Adversarial Perturbations via Gated Batch Normalization
Figure 3 for Towards Defending Multiple Adversarial Perturbations via Gated Batch Normalization
Figure 4 for Towards Defending Multiple Adversarial Perturbations via Gated Batch Normalization
* 28 pages 
  Access Paper or Ask Questions

Adversarial Examples for $k$-Nearest Neighbor Classifiers Based on Higher-Order Voronoi Diagrams


Nov 19, 2020
Chawin Sitawarin, Evgenios M. Kornaropoulos, Dawn Song, David Wagner

Figure 1 for Adversarial Examples for $k$-Nearest Neighbor Classifiers Based on Higher-Order Voronoi Diagrams
Figure 2 for Adversarial Examples for $k$-Nearest Neighbor Classifiers Based on Higher-Order Voronoi Diagrams
Figure 3 for Adversarial Examples for $k$-Nearest Neighbor Classifiers Based on Higher-Order Voronoi Diagrams
Figure 4 for Adversarial Examples for $k$-Nearest Neighbor Classifiers Based on Higher-Order Voronoi Diagrams

  Access Paper or Ask Questions

Enabling Efficient Cyber Threat Hunting With Cyber Threat Intelligence


Oct 26, 2020
Peng Gao, Fei Shao, Xiaoyuan Liu, Xusheng Xiao, Zheng Qin, Fengyuan Xu, Prateek Mittal, Sanjeev R. Kulkarni, Dawn Song

Figure 1 for Enabling Efficient Cyber Threat Hunting With Cyber Threat Intelligence
Figure 2 for Enabling Efficient Cyber Threat Hunting With Cyber Threat Intelligence
Figure 3 for Enabling Efficient Cyber Threat Hunting With Cyber Threat Intelligence
Figure 4 for Enabling Efficient Cyber Threat Hunting With Cyber Threat Intelligence

  Access Paper or Ask Questions

Language Models are Open Knowledge Graphs


Oct 22, 2020
Chenguang Wang, Xiao Liu, Dawn Song

Figure 1 for Language Models are Open Knowledge Graphs
Figure 2 for Language Models are Open Knowledge Graphs
Figure 3 for Language Models are Open Knowledge Graphs
Figure 4 for Language Models are Open Knowledge Graphs
* 30 pages, 32 figures, 3 tables 
  Access Paper or Ask Questions

F2ED-Learning: Good Fences Make Good Neighbors


Oct 02, 2020
Lun Wang, Qi Pang, Shuai Wang, Dawn Song

Figure 1 for F2ED-Learning: Good Fences Make Good Neighbors
Figure 2 for F2ED-Learning: Good Fences Make Good Neighbors
Figure 3 for F2ED-Learning: Good Fences Make Good Neighbors

  Access Paper or Ask Questions

Model-Agnostic Round-Optimal Federated Learning via Knowledge Transfer


Oct 02, 2020
Qinbin Li, Bingsheng He, Dawn Song

Figure 1 for Model-Agnostic Round-Optimal Federated Learning via Knowledge Transfer
Figure 2 for Model-Agnostic Round-Optimal Federated Learning via Knowledge Transfer
Figure 3 for Model-Agnostic Round-Optimal Federated Learning via Knowledge Transfer
Figure 4 for Model-Agnostic Round-Optimal Federated Learning via Knowledge Transfer

  Access Paper or Ask Questions

Measuring Massive Multitask Language Understanding


Sep 21, 2020
Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, Jacob Steinhardt

Figure 1 for Measuring Massive Multitask Language Understanding
Figure 2 for Measuring Massive Multitask Language Understanding
Figure 3 for Measuring Massive Multitask Language Understanding
Figure 4 for Measuring Massive Multitask Language Understanding
* The test and code is available at https://github.com/hendrycks/test 
  Access Paper or Ask Questions

A Principled Approach to Data Valuation for Federated Learning


Sep 14, 2020
Tianhao Wang, Johannes Rausch, Ce Zhang, Ruoxi Jia, Dawn Song

Figure 1 for A Principled Approach to Data Valuation for Federated Learning
Figure 2 for A Principled Approach to Data Valuation for Federated Learning
Figure 3 for A Principled Approach to Data Valuation for Federated Learning
Figure 4 for A Principled Approach to Data Valuation for Federated Learning

  Access Paper or Ask Questions

Compositional Generalization via Neural-Symbolic Stack Machines


Aug 15, 2020
Xinyun Chen, Chen Liang, Adams Wei Yu, Dawn Song, Denny Zhou

Figure 1 for Compositional Generalization via Neural-Symbolic Stack Machines
Figure 2 for Compositional Generalization via Neural-Symbolic Stack Machines
Figure 3 for Compositional Generalization via Neural-Symbolic Stack Machines
Figure 4 for Compositional Generalization via Neural-Symbolic Stack Machines

  Access Paper or Ask Questions

Aligning AI With Shared Human Values


Aug 05, 2020
Dan Hendrycks, Collin Burns, Steven Basart, Andrew Critch, Jerry Li, Dawn Song, Jacob Steinhardt

Figure 1 for Aligning AI With Shared Human Values
Figure 2 for Aligning AI With Shared Human Values
Figure 3 for Aligning AI With Shared Human Values
Figure 4 for Aligning AI With Shared Human Values

  Access Paper or Ask Questions

Synthesize, Execute and Debug: Learning to Repair for Neural Program Synthesis


Jul 16, 2020
Kavi Gupta, Peter Ebert Christensen, Xinyun Chen, Dawn Song

Figure 1 for Synthesize, Execute and Debug: Learning to Repair for Neural Program Synthesis
Figure 2 for Synthesize, Execute and Debug: Learning to Repair for Neural Program Synthesis
Figure 3 for Synthesize, Execute and Debug: Learning to Repair for Neural Program Synthesis
Figure 4 for Synthesize, Execute and Debug: Learning to Repair for Neural Program Synthesis

  Access Paper or Ask Questions

The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization


Jun 29, 2020
Dan Hendrycks, Steven Basart, Norman Mu, Saurav Kadavath, Frank Wang, Evan Dorundo, Rahul Desai, Tyler Zhu, Samyak Parajuli, Mike Guo, Dawn Song, Jacob Steinhardt, Justin Gilmer

Figure 1 for The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization
Figure 2 for The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization
Figure 3 for The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization
Figure 4 for The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization
* Datasets, code, and models available at https://github.com/hendrycks/imagenet-r 
  Access Paper or Ask Questions

Towards practical differentially private causal graph discovery


Jun 15, 2020
Lun Wang, Qi Pang, Dawn Song

Figure 1 for Towards practical differentially private causal graph discovery
Figure 2 for Towards practical differentially private causal graph discovery
Figure 3 for Towards practical differentially private causal graph discovery

  Access Paper or Ask Questions

Imitation Attacks and Defenses for Black-box Machine Translation Systems


Apr 30, 2020
Eric Wallace, Mitchell Stern, Dawn Song

Figure 1 for Imitation Attacks and Defenses for Black-box Machine Translation Systems
Figure 2 for Imitation Attacks and Defenses for Black-box Machine Translation Systems
Figure 3 for Imitation Attacks and Defenses for Black-box Machine Translation Systems
Figure 4 for Imitation Attacks and Defenses for Black-box Machine Translation Systems

  Access Paper or Ask Questions

Pretrained Transformers Improve Out-of-Distribution Robustness


Apr 16, 2020
Dan Hendrycks, Xiaoyuan Liu, Eric Wallace, Adam Dziedzic, Rishabh Krishnan, Dawn Song

Figure 1 for Pretrained Transformers Improve Out-of-Distribution Robustness
Figure 2 for Pretrained Transformers Improve Out-of-Distribution Robustness
Figure 3 for Pretrained Transformers Improve Out-of-Distribution Robustness
Figure 4 for Pretrained Transformers Improve Out-of-Distribution Robustness
* ACL 2020 
  Access Paper or Ask Questions

Anomalous Instance Detection in Deep Learning: A Survey


Mar 16, 2020
Saikiran Bulusu, Bhavya Kailkhura, Bo Li, Pramod K. Varshney, Dawn Song

Figure 1 for Anomalous Instance Detection in Deep Learning: A Survey
Figure 2 for Anomalous Instance Detection in Deep Learning: A Survey
Figure 3 for Anomalous Instance Detection in Deep Learning: A Survey
Figure 4 for Anomalous Instance Detection in Deep Learning: A Survey

  Access Paper or Ask Questions

Proceedings of the AAAI-20 Workshop on Intelligent Process Automation (IPA-20)


Feb 26, 2020
Dell Zhang, Andre Freitas, Dacheng Tao, Dawn Song


  Access Paper or Ask Questions