Dimitris Tsipras

<!doctype html> <html lang="en"> <head>  <meta charset="utf-8"> <meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no">  <link rel="stylesheet" href="https://stackpath.bootstrapcdn.com/bootstrap/4.1.3/css/bootstrap.min.css" integrity="sha384-MCw98/SFnGE8fJT3GXwEOngsV7Zt27NXFoaoApmYm81iuXoPkFOJwJ8ERdknLPMO" crossorigin="anonymous">  <link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/4.7.0/css/font-awesome.min.css"> <link rel="stylesheet" href="https://cdn.rawgit.com/jpswalsh/academicons/master/css/academicons.min.css"> <style> .nonselect, #all-btn { display: none } h3 { line-height: .8 } </style> <title>Dimitris Tsipras</title> </head> <body> <div class="container mt-5"> <div class="row"> <div class="col-md-5 offset-md-2"> <a name="aboaut"><h3>Dimitris Tsipras</h3></a> <p> <br> I am a research scientist at OpenAI. <br><br> Before that, I was a postdoc at Stanford CS, mentored by <a href="https://cs.stanford.edu/~pliang/">Percy Liang</a> and <a href="https://theory.stanford.edu/~valiant/">Greg Valiant</a>, working on LLMs and in-context learning. <br><br> I did my PhD in CS at MIT with <a href="http://people.csail.mit.edu/madry">Aleksander M膮dry</a>, working on optimization and ML robustness. <br><br> I got my undergrad degree in ECE from NTUA, Greece, where I did my diploma thesis with <a href="https://www.softlab.ntua.gr/~fotakis">Dimitris Fotakis</a>. <br><br> </p> </div> <div class="col-md-3"> <img src="prof.jpg" alt="my face" class="img-fluid rounded"> <div class="row pt-3"> <div class="col-2"> <a href="http://github.com/dtsip"> <i class="fa fa-github" style="font-size:30px; color: black"></i> </a> </div> <div class="col-1"> <a href="https://scholar.google.com/citations?user=26eh1jAAAAAJ"> <i class="ai ai-google-scholar" style="font-size:28px; padding-top:1px; color: #4285f4"></i> </a> </div> </div> </div> </div> </div> <div class="container pt-3"> <div class="row"> <div class="col-md-8 offset-md-2"> <h3> Selected papers <font size="-1"> <a id="selected-btn" href="javascript:void(0)">Show all</a> <a id="all-btn" href="javascript:void(0)">Show selected</a> </font><br> <font size="-1">(* means equal contribution)</font> </h3> <br> <div class="nonselect"> <a href="https://arxiv.org/abs/2211.09110"> Holistic Evaluation of Language Models</a><br> Percy Liang, Rishi Bommasani, Tony Lee, Dimitris Tsipras, Dilara Soylu, Michihiro Yasunaga, Yian Zhang, Deepak Narayanan, Yuhuai Wu, Ananya Kumar, Benjamin Newman, Binhang Yuan, Bobby Yan, Ce Zhang, Christian Cosgrove, Christopher D. Manning, Christopher R茅, Diana Acosta-Navas, Drew A. Hudson, Eric Zelikman, Esin Durmus, Faisal Ladhak, Frieda Rong, Hongyu Ren, Huaxiu Yao, Jue Wang, Keshav Santhanam, Laurel Orr, Lucia Zheng, Mert Yuksekgonul, Mirac Suzgun, Nathan Kim, Neel Guha, Niladri Chatterji, Omar Khattab, Peter Henderson, Qian Huang, Ryan Chi, Sang Michael Xie, Shibani Santurkar, Surya Ganguli, Tatsunori Hashimoto, Thomas Icard, Tianyi Zhang, Vishrav Chaudhary, William Wang, Xuechen Li, Yifan Mai, Yuhui Zhang, Yuta Koreeda<br> TMLR 2023 <br> <a href="https://crfm.stanford.edu/helm/lite/latest/">website</a>, <a href="https://github.com/stanford-crfm/helm">code</a> <br><br> </div> <div> <a href="https://arxiv.org/abs/2208.01066"> What Can Transformers Learn In-Context? A Case Study of Simple Function Classes</a><br> Shivam Garg*, Dimitris Tsipras*, Percy Liang, Gregory Valiant<br> NeurIPS 2022<br> <a href="https://github.com/dtsip/in-context-learning">code</a> <br><br> </div> <div class="nonselect"> <a href="https://arxiv.org/abs/2112.01008"> Editing a Classifier by Rewriting Its Prediction Rules</a><br> Shibani Santurkar*, Dimitris Tsipras*, Mahi Elango, David Bau, Antonio Torralba, Aleksander M膮dry<br> NeurIPS 2021 <br> <a href="http://gradientscience.org/editing">blog post</a>, <a href="https://github.com/MadryLab/EditingClassifiers">code</a> <br><br> </div> <div class="nonselect"> <a href="https://arxiv.org/abs/2110.08220"> Combining Diverse Feature Priors</a><br> Saachi Jain*, Dimitris Tsipras*, Aleksander M膮dry<br> ICML 2022 <br> <a href="http://gradientscience.org/copriors">blog post</a>, <a href="https://github.com/MadryLab/copriors">code</a> <br><br> </div> <div class="nonselect"> <a href="https://arxiv.org/abs/2012.10544"> Dataset Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses</a><br> Micah Goldblum, Dimitris Tsipras, Chulin Xie, Xinyun Chen, Avi Schwarzschild, Dawn Song, Aleksander Madry, Bo Li, Tom Goldstein <br> TPAMI 2022 <br><br> </div> <div class=nonselect> <a href="https://arxiv.org/abs/2008.04859"> BREEDS: Benchmarks for Subpopulation Shift</a><br> Shibani Santurkar*, Dimitris Tsipras*, Aleksander M膮dry<br> ICLR 2021 <br> <a href="http://gradientscience.org/breeds">blog post</a>, <a href="https://github.com/MadryLab/BREEDS-Benchmarks">benchmarks</a> <br><br> </div> <div> <a href="https://arxiv.org/abs/2005.11295">From ImageNet to Image Classification: Contextualizing Progress on Benchmarks</a><br> Dimitris Tsipras*, Shibani Santurkar*, Logan Engstrom, Andrew Ilyas, Aleksander M膮dry<br> ICML 2020 <br> <a href="http://gradientscience.org/benchmarks">blog post</a>, <a href="https://github.com/MadryLab/ImageNetMultiLabel">ImageNet annotations</a> <br><br> </div> <div class="nonselect"> <a href="https://arxiv.org/abs/2005.09619">Identifying Statistical Bias in Dataset Replication</a><br> Logan Engstrom*, Andrew Ilyas*, Shibani Santurkar, Dimitris Tsipras, Jacob Steinhardt, Aleksander M膮dry<br> ICML 2020 <br> <a href="http://gradientscience.org/data_rep_bias/">blog post</a>, <a href="https://github.com/MadryLab/dataset-replication-analysis">code</a> <br><br> </div> <div class="nonselect"> <a href="https://arxiv.org/abs/1906.09453">Image Synthesis with a Single (Robust) Classifier</a><br> Shibani Santurkar*, Dimitris Tsipras*, Brandon Tran*, Andrew Ilyas*, Logan Engstrom*, Aleksander M膮dry<br> NeurIPS 2019 <br> <a href="http://gradientscience.org/robust_apps/">blog post</a>, <a href="http://git.io/robust-apps">code/notebooks</a>, <a href="http://bit.ly/robustness_demo">demo</a> <br><br> </div> <div class="nonselect"> <a href="https://arxiv.org/abs/1906.00945">Learning Perceptually-Aligned Representations via Adversarial Robustness</a><br> Logan Engstrom*, Andrew Ilyas*, Shibani Santurkar*, Dimitris Tsipras*, Brandon Tran*, Aleksander M膮dry<br> <a href="http://gradientscience.org/robust_reps">blog post</a>, <a href="http://git.io/robust-reps">code/notebooks</a> <br><br> </div> <div> <a href="https://arxiv.org/abs/1905.02175">Adversarial Examples Are Not Bugs, They Are Features</a><br> Andrew Ilyas*, Shibani Santurkar*, Dimitris Tsipras*, Logan Engstrom*, Brandon Tran, Aleksander M膮dry<br> NeurIPS 2019<br> <a href="http://gradientscience.org/adv">blog post</a>, <a href="http://git.io/adv-datasets">datasets</a> <br><br> </div> <div class="nonselect"> <a href="https://arxiv.org/abs/1902.06705">On Evaluating Adversarial Robustness</a><br> Nicholas Carlini, Anish Athalye, Nicolas Papernot, Wieland Brendel, Jonas Rauber, Dimitris Tsipras, Ian Goodfellow, Aleksander M膮dry, Alexey Kurakin<br><br> </div> <div class="nonselect"> <a href="https://arxiv.org/abs/1912.02771">Label-Consistent Backdoor Attacks</a><br> Alexander Turner, Dimitris Tsipras, Aleksander M膮dry <br><br> </div> <div class="nonselect"> <a href="https://arxiv.org/abs/1811.02553">A Closer Look at Deep Policy Gradients</a><br> Andrew Ilyas*, Logan Engstrom*, Shibani Santurkar, Dimitris Tsipras, Firdaus Janoos, Larry Rudolph, Aleksander M膮dry<br> ICLR 2020<br> blog post: <a href="https://gradientscience.org/policy_gradients_pt1">part 1</a>, <a href="https://gradientscience.org/policy_gradients_pt2">part 2</a>, <a href="https://gradientscience.org/policy_gradients_pt3">part 3</a><br><br> </div> <div class="nonselect"> <a href="https://arxiv.org/abs/2005.12729">Implementation Matters in Deep Policy Gradients: A Case Study on PPO and TRPO</a><br> Logan Engstrom*, Andrew Ilyas*, Shibani Santurkar, Dimitris Tsipras, Firdaus Janoos, Larry Rudolph, Aleksander M膮dry<br> ICLR 2020<br> <a href="https://github.com/MadryLab/implementation-matters">code</a><br><br> </div> <div> <a href="https://arxiv.org/abs/1805.12152">Robustness May Be at Odds with Accuracy</a><br> Dimitris Tsipras*, Shibani Santurkar*, Logan Engstrom*, Alexander Turner, Aleksander M膮dry<br> ICLR 2019<br><br> </div> <div> <a href="https://arxiv.org/abs/1805.11604">How Does Batch Normalization Help Optimization?</a><br> Shibani Santurkar*, Dimitris Tsipras*, Andrew Ilyas*, Aleksander M膮dry<br> NeurIPS 2018<br> <a href="https://gradientscience.org/batchnorm">blog post</a>, <a href="https://www.youtube.com/watch?v=ZOabsYbmBRM">3-minute video</a><br><br> </div> <div class="nonselect"> <a href="https://arxiv.org/abs/1804.11285">Adversarially Robust Generalization Requires More Data</a><br> Ludwig Schmidt, Shibani Santurkar, Dimitris Tsipras, Kunal Talwar, Aleksander M膮dry<br> NeurIPS 2018<br><br> </div> <div class="nonselect"> <a href="https://arxiv.org/abs/1712.02779">Exploring the Landscape of Spatial Robustness</a><br> Logan Engstrom*, Brandon Tran*, Dimitris Tsipras*, Ludwig Schmidt, Aleksander M膮dry<br> ICML 2019<br> <a href="https://github.com/MadryLab/adversarial_spatial">code</a><br><br> </div> <div> <a href="https://arxiv.org/abs/1706.06083">Towards Deep Learning Models Resistant to Adversarial Attacks</a><br> (伪-尾 order) Aleksander M膮dry, Aleksandar Makelov, Ludwig Schmidt, Dimitris Tsipras, Adrian Vladu<br> ICLR 2018<br> blog post: <a href="https://gradientscience.org/robust_opt_pt1">part 1</a> and <a href="https://gradientscience.org/robust_opt_pt2/">part 2</a>, <a href="https://github.com/MadryLab/mnist_challenge">MNIST Challenge</a>, <a href="https://github.com/MadryLab/cifar10_challenge">CIFAR10 Challenge</a><br><br> </div> <div class="nonselect"> <a href="https://arxiv.org/abs/1704.02310">Matrix Scaling and Balancing via Box Constrained Newton's Method and Interior Point Methods</a><br> (伪-尾 order) Michael B. Cohen, Aleksander M膮dry, Dimitris Tsipras, Adrian Vladu<br> FOCS 2017<br><br> </div> <div class="nonselect"> <a href="pdfs/FTTZ15.pdf">Efficient Money Burning in General Domains</a><br> (伪-尾 order) Dimitris Fotakis, Dimitris Tsipras, Christos Tzamos, Manolis Zampetakis <br> SAGT 2015<br> Invited to special issue of Theory of Computing Systems<br><br> </div> </div> </div>   <script src="https://code.jquery.com/jquery-3.3.1.slim.min.js" integrity="sha384-q8i/X+965DzO0rT7abK41JStQIAqVgRVzpbzo5smXKp4YfRvH+8abtTE1Pi6jizo" crossorigin="anonymous"></script> <script src="https://cdnjs.cloudflare.com/ajax/libs/popper.js/1.14.3/umd/popper.min.js" integrity="sha384-ZMP7rVo3mIykV+2+9J3UJ46jBk0WLaUAdn689aCwoqbBJiSnjAK/l8WvCWPIPm49" crossorigin="anonymous"></script> <script src="https://stackpath.bootstrapcdn.com/bootstrap/4.1.3/js/bootstrap.min.js" integrity="sha384-ChfqqxuZUCnJSK3+MXmPNIyE6ZbWh2IMqE241rYiqJxyMiZ6OW/JmZQ5stwEULTy" crossorigin="anonymous"></script> <script> $("#selected-btn").click(function(){ $(".nonselect").show(); $("#all-btn").show(); $("#selected-btn").hide(); }); $("#all-btn").click(function(){ $(".nonselect").hide(); $("#all-btn").hide(); $("#selected-btn").show(); }); </script> </body> </html>

CINXE.COM

Dimitris Tsipras