Dan Hendrycks

I am a third-year PhD student in computer science at UC Berkeley where I am advised by Dawn Song and Jacob Steinhardt. I am interested in Trustworthy Machine Learning. In 2018 I received my BS from UChicago. My research is supported by the NSF GRFP and The Open Philanthropy Project. My contributions include the GELU activation function, the out-of-distribution detection baseline, and distribution shift benchmarks. hendrycks ατ berkeley.edu Research Updates (Twitter) | Google Scholar | Semantic Scholar

Works

• Measuring Massive Multitask Language Understanding
  Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, Jacob Steinhardt
  [arXiv] [code]


• Aligning AI With Shared Human Values
  Dan Hendrycks, Collin Burns, Steven Basart, Andrew Critch, Jerry Li, Dawn Song, Jacob Steinhardt
  [arXiv] [code]


• The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization
  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
  [arXiv] [code]


• Pretrained Transformers Improve Out-of-Distribution Robustness
  Dan Hendrycks*, Xiaoyuan Liu*, Eric Wallace, Adam Dziedzic, Rishabh Krishnan, Dawn Song
  ACL 2020
  [arXiv] [code]


• AugMix: A Simple Data Processing Method to Improve Robustness and Uncertainty
  Dan Hendrycks*, Norman Mu*, Ekin D. Cubuk, Barret Zoph, Justin Gilmer, Balaji Lakshminarayanan
  ICLR 2020
  [arXiv] [code]


• A Benchmark for Anomaly Segmentation
  Dan Hendrycks*, Steven Basart*, Mantas Mazeika, Mohammadreza Mostajabi, Jacob Steinhardt, Dawn Song
  [arXiv] [code]


• Natural Adversarial Examples
  Dan Hendrycks, Kevin Zhao, Steven Basart, Jacob Steinhardt, Dawn Song
  [arXiv] [code]


• Using Self-Supervised Learning Can Improve Model Robustness and Uncertainty
  Dan Hendrycks, Mantas Mazeika*, Saurav Kadavath*, Dawn Song
  NeurIPS 2019
  [arXiv] [code]


• Testing Robustness Against Unforeseen Adversaries
  Daniel Kang*, Yi Sun*, Dan Hendrycks, Tom Brown, Jacob Steinhardt
  [arXiv] [code]


• Adversarial Example Researchers Need to Expand What is Meant by ‘Robustness’
  Justin Gilmer, Dan Hendrycks
  Distill 2019
  


• Using Pre-Training Can Improve Model Robustness and Uncertainty
  Dan Hendrycks, Kimin Lee, Mantas Mazeika
  ICML 2019, the first paper in the literature to show adversarial training with more data works
  [arXiv] [code]


• Deep Anomaly Detection with Outlier Exposure
  Dan Hendrycks, Mantas Mazeika, Thomas Dietterich
  ICLR 2019
  [arXiv] [code]


• Benchmarking Neural Network Robustness to Common Corruptions and Perturbations
  Dan Hendrycks, Thomas Dietterich
  ICLR 2019
  [arXiv] [code]


• Using Trusted Data to Train Deep Networks on Labels Corrupted by Severe Noise
  Dan Hendrycks*, Mantas Mazeika*, Duncan Wilson, Kevin Gimpel
  NeurIPS 2018
  [arXiv] [code]


• Open Category Detection with PAC Guarantees
  Si Liu, Risheek Garrepalli, Thomas G. Dietterich, Alan Fern, Dan Hendrycks
  ICML 2018
  [arXiv] [code]


• A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks
  Dan Hendrycks, Kevin Gimpel
  ICLR 2017, in which we coin "out-of-distribution detection"
  [arXiv] [code]


• Early Methods for Detecting Adversarial Images
  Dan Hendrycks, Kevin Gimpel
  ICLR 2017 Workshop, the first paper in the literature on adversarial example detection.
  [arXiv]


• Gaussian Error Linear Units (GELUs)
  Dan Hendrycks, Kevin Gimpel
  Technical Report, in which we propose the SiLU (x*σ(x)) and the GELU (x*Φ(x)). The GELU is included in the PyTorch and Tensorflow libraries, used in BERT and GPT, and is one of the most used activation functions.
  [arXiv]

Service