Photo of Logan Engstrom

Logan Engstrom

Email: engstrom@mit.edu
GitHub: @lengstrom
Google Scholar: here
CV/Resume: here

About

I'm a first year PhD student at MIT researching machine learning. I am advised by Aleksander Mądry. I am most interested in robust/reliable machine learning and understanding ML foundations. I previously worked at Two Sigma (summer 2018), Google Brain (summer 2017) and Apple (summer 2016).

Outside of research I contribute to open source, and at MIT I play for the ESP intramural flag football and dodgeball teams. I used to help organize HackMIT.

Publications

  1. Andrew Ilyas*, Shibani Santurkar*, Dimitris Tsipras*, Logan Engstrom*, Brandon Tran, and Aleksander Madry. Adversarial examples are not bugs, they are features. NeurIPS 2019.

  2. Shibani Santurkar*, Dimitris Tsipras*, Brandon Tran*, Andrew Ilyas*, Logan Engstrom*, and Aleksander Madry. Computer vision with a single (robust) classifier.. NeurIPS 2019.

  3. Dimitris Tsipras*, Shibani Santurkar*, Logan Engstrom*, Alexander Turner, and Aleksander Mądry. Robustness May Be at Odds with Accuracy. ICLR 2019.

  4. Andrew Ilyas*, Logan Engstrom*, Ludwig Schmidt, and Aleksander Mądry. Prior Convictions: Black-Box Adversarial Attacks with Bandits and Priors. ICLR 2019.

  5. Logan Engstrom*, Brandon Tran*, Dimitris Tsipras*, Ludwig Schmidt, and Aleksander Mądry. A Rotation and a Translation Suffice: Fooling CNNs with Simple Transformations. ICML 2019.

  6. Logan Engstrom*, Andrew Ilyas*, and Anish Athalye*. Evaluating and Understanding the Robustness of Adversarial Logit Pairing. NIPS 2018 Machine Learning and Computer Security Workshop.

  7. Andrew Ilyas*, Logan Engstrom*, Anish Athalye*, and Jessy Lin*. Black-box Adversarial Attacks with Limited Queries and Information. ICML 2018.

  8. Anish Athalye*, Logan Engstrom*, Andrew Ilyas*, and Kevin Kwok. Synthesizing Robust Adversarial Examples. ICML 2018, Demo at NeurIPS 2017 Machine Learning and Computer Security Workshop.

  9. Daniel Kang, Richard Sherwood, Amira Barkal, Tatsunori Hashimoto, Logan Engstrom, and David Gifford. DNase-capture Reveals Differential Transcription Factor Binding modalities. PloS one, 2017.

* indicates equal contribution

Preprints

  1. Andrew Ilyas*, Logan Engstrom*, Shibani Santurkar, Dimitris Tsipras, Firdaus Janoos, Larry Rudolph, and Aleksander Mądry. Are Deep Policy Gradient Algorithms Truly Policy Gradient Algorithms?. 2018.

  2. Logan Engstrom*, Andrew Ilyas*, Shibani Santurkar*, Dimitris Tsipras*, Brandon Tran*, and Aleksander Madry. Learning perceptually-aligned representations via adversarial robustness. 2019.

* indicates equal contribution

Projects