Sarah H. Cen

I am a Ph.D. student (MIT EECS) advised by Prof. Devavrat Shah and Prof. Aleksander Mądry. My research lies at the intersection of machine learning theory and AI ethics. I am interested in the design of data-driven algorithms as well as their governance, and my work uses tools from statistics, causal inference, and economics.

Recently, I have written on how to audit social media algorithms, the estimation of counterfactual potential outcomes under spillover effects, how competing for resources under uncertainty affects long-term outcomes, and the individual-level rights of data-driven decision subjects. Previously, I worked in intelligent transportation, communication networks, reinforcement learning, and robotics. During my master's, I performed research on autonomous vehicles (a.k.a. self-driving cars) with Prof. Paul Newman at the Unversity of Oxford. As an undergraduate, I studied control & decision systems with Prof. Naomi Leonard at Princeton University.

Research

To see my research experiences, click here.

Publications and Pre-prints

*First-author contribution.

1. A. Agarwal, S. H. Cen*, D. Shah, and C. L. Yu. "Network Synthetic Interventions: A Causal Framework for Panel Data with Network Interference." Pre-print. 2022.
2. S. H. Cen* and M. Raghavan. "The Right to be an Exception to a Data-Driven Rule." Pre-print. 2022.
3. J. Perolat*, B. De Vylder*, D. Hennes*, [...], S. H. Cen, et al. "Mastering the game of Stratego with model-free multiagent reinforcement learning." In Science. 2022.
4. S. H. Cen* and D. Shah. "Regret, stability, and fairness in matching markets with bandit learners." In Proceedings of the 25th International Conference on Artificial Intelligence and Statistics (AISTATS), 2022.
5. S. H. Cen* and D. Shah. "Regulating algorithmic filtering on social media." Spotlight paper. In Proceedings of the Thirty-fifth Conference on Neural Information Processing Systems (NeurIPS), 2021.
6. S. H. Cen* and P. Newman. "Radar-only ego-motion estimation in difficult settings via graph matching." In Proceedings of the IEEE Intl. Conf. on Robotics and Automation (ICRA), 2019.
7. R. Weston*, S. H. Cen, P. Newman, and I. Posner. "Probably Unknown: Deep Inverse Sensor Modelling in Radar." In Proceedings of the IEEE Intl. Conf. on Robotics and Automation (ICRA), 2019.
8. S. H. Cen* and P. Newman. "Precise Ego-Motion Estimation with Millimeter-Wave Radar under Diverse and Challenging Conditions." In Proceedings of the IEEE Intl. Conf. on Robotics and Automation (ICRA), 2018.
9. S. H. Cen*, V. Srivastava, and N. E. Leonard. "On robustness and leadership in Markov switching consensus networks." In Proceedings of the IEEE Conference on Decision and Control (CDC), 2017.

---

Selected Recent Talks

• The Right to be an Exception to a Data-Driven Rule - manuscript | slides
  Presented at the Privacy Law Scholars Conference (June 2022), the Harvard Ethical Issues in Computing & AI Conference (June 2022), and the MIT AI Ethics Reading Group.
• Regulating Algorithmic Filtering on Social Media - manuscript | slides | poster
  Presented as a Spotlight at NeurIPS (December 2021), LIDS Student Conference (January 2022) and received Best Student Presentation, LIDS Stats & Tea Talk (April 2021), and Stanford GSB Rising Scholars Conference (October 2020).
• Regret, stability, and fairness in matching markets with bandit learners - manuscript | slides | poster
  Presented as poster at AISTATS (March 2022) and as a LIDS Stats & Tea Talk (May 2021).
• A Game-Theoretic Perspective on Trust in Recommendation - slides
  Presented as a Contributed Talk in the ICML Workshop on Responsible Decision Making in Dynamic Environments (July 2022) and at the Simon's Institute for Theory of Computing's Workshop on AI & Humanity (July 2022).
• Why Social Media Platforms Shape What You Think - slides
  Presented at INFORMS (October 2022).
• Making Online Platforms Fit For Humans
  Presented as a Thriving Star of AI at MIT (May 2022).

---

Media and Videos

• August 2022: Lance Eliot wrote a column in Forbes titled "AI Ethics Fighting Passionately For Your Right To Be An Exception" based on my paper "The Right to be an Exception to a Data-Driven Rule" with Manish Raghavan!
• July 2022: I was honored to give a talk at the Simons Institute for the Theory of Computing on "The Design and Governance of Human-Facing Algorithms" at the workshop for AI and Humanity!
• May 2022: I was lucky enough to have MIT News feature my research in the article "Living better with algorithms" by Grace Chua!
• May 2022: I am grateful to have been selected as a Thriving Star of AI at MIT!
• December 2021: My paper on "Regulating algorithmic filtering on social media" written with Devavrat Shah was featured as a Spotlight at Neurips 2021!
• May 2019: My paper on "Radar-only ego-motion estimation in difficult settings via graph matching" written with Paul Newman was presented at ICRA 2019!
• May 2018: My paper on "Precise Ego-Motion Estimation with Millimeter-Wave Radar under Diverse and Challenging Conditions" written with Paul Newman was presented at ICRA 2018!

Education

• Inference & Information  |
• Probability Theory  |
• Algorithms for Inference  |
• Linear Optimization  |
• Mathematical Statistics  |
• Discrete Probability & Stochastic Processes  |
• Individual Risk  |
• Artificial Intelligence  |
• Microeconomic Theory  |
• Game Theory  |
• Automatic Control Systems  |
• Big Data  |
• Ethics & Fairness in Data-Driven Decision-Making  |
• Law & Ethics of AI  |
• Urban Sociology  |
• Race & Ethnicity  |
• History of Poverty  |
• Political Theory  |
• Human Rights

Experience

Honors

More information

Leadership positions

---

Teaching

"The Science Policy Bootcamp is a 5-day short course, offered during MIT's Independent Activities Period in January, designed to introduce participants to the 'nuts and bolts' of science policy making. The course provides an opportunity for young scientists and engineers interested in science policy issues to increase their understanding about and practical involvement with science policy. The bootcamp serves to both expose participants to the fundamental structure and dynamics of science policy and inform them of routes into a policy experience or career."

"Introduction to principles of Bayesian and non-Bayesian statistical inference. Hypothesis testing and parameter estimation, sufficient statistics; exponential families. EM agorithm. Log-loss inference criterion, entropy and model capacity. Kullback-Leibler distance and information geometry. Asymptotic analysis and large deviations theory. Model order estimation; nonparametric statistics. Computational issues and approximation techniques; Monte Carlo methods. Selected topics such as universal inference and learning, and universal features and neural networks."

---

Programming Languages

• C  |
• C++  |
• Python  |
• MATLAB  |
• Java  |
• HTML  |
• JS  |
• UNIX