Sarah Scheffler

Seeking Ph.D. students to apply Dec 2024

I am seeking Ph.D. student applicants for the next academic year (applications due December 11, 2024) I am especially seeking students who are interested in working on:

Content moderation in encrypted environments
Cryptographic verification (e.g. zero-knowledge proofs)
Privacy and data confidentiality in robotic applications
The intersection of (mostly U.S.) law with specific parts of cryptography/privacy (especially self-incrimination, surveillance, and sometimes copyright)
Finding the "right" setting of privacy in cryptographic privacy-preserving applications (e.g. when is it okay to link a record?)
Anyone with interest/background in cryptography or privacy who is interested in applied work that is relevant to policy, regulation, law, or general society

If any of those sounds like your jam, please apply to the Carnegie Mellon University Ph.D. programs in either the Engineering and Public Policy Ph.D. Program (in the College of Engineering) or the Societal Computing Ph.D. program (in the School of Computer Science).

Please submit your application and write my name somewhere in it; I have very limited bandwidth to respond to prospective student emails. I typically interview interested students in the early spring..

About me

I am an assistant professor at CyLab, Carnegie Mellon University's Security and Privacy Institute, studying at the intersection of cryptography and policy. I am jointly appointed between Software and Societal Systems (in the School of Computer Science) and Engineering and Public Policy (in the College of Engineering). If you are interested in joining me as a student in the Fall of 2024, please first poke around this page to see what kind of work I do, and apply to either department at CMU. (I can sometimes advise students outside of S3D or EPP; if you applied to a different department, please email me.) My ideal student has a strong background or interest in cryptography and its intersection with societal, legal, and policy issues.

I am a studying applied cryptographer working at the intersection of cryptography, privacy, policy, and law. My interdisciplinary work includes policy and technical analysis of end-to-end encrypted content moderation, compelled decryption, and privacy-preserving computation as applied to journalism, age verification, and more. I also do "pure" applied cryptography, including work on zero-knowledge proofs, multi-party computation, private set intersection, and hash combiners.

Formerly, I was a postdoctoral research associate at MIT's Internet Policy Research Initiative, and before that I was a postdoctoral research associate at Princeton University's Center for Information Technology Policy. I obtained my Ph.D. from Boston University in 2021, advised by Prof. Mayank Varia. During my time at BU, I was a Ph.D. student in the BUsec group, I organized the BUsec Seminar for security, cryptography, and privacy, as well as the Multi-Party Computation Reading Group, and I was an active member of the Cyber Security, Law, and Society Alliance.

Current Students

Shuang Liu

Peer-reviewed Publications

When Anti-Fraud Laws become a Barrier to Computer Science Research
Madelyne Xiao, Andrew Sellars, Sarah Scheffler
ACM CS/Law 2025 (forthcoming)

Group Moderation under End-to-End Encryption
Sarah Scheffler, Jonathan Mayer
ACM CS/Law 2024

SoK: Content Moderation for End-to-End Encryption
Sarah Scheffler, Jonathan Mayer
PoPETS 2023

Public Verification for Private Hash Matching
Sarah Scheffler, Anunay Kulshrestha, Jonathan Mayer
IEEE S&P 2023

Can the government compel decryption? Don't trust --- verify
Aloni Cohen, Sarah Scheffler, Mayank Varia
ACM CS/Law 2022

Formalizing Human Ingenuity: A Quantitative Framework for Copyright Law's Substantial Similarity
Eran Tromer, Sarah Scheffler, Mayank Varia
ACM CS/Law 2022

TurboIKOS: Improved Non-interactive Zero Knowledge with Sublinear Memory
Yaron Gvili, Julie Ha, Sarah Scheffler, Mayank Varia, Ziling Yang, Xinyuan Zhang
ACNS 2021

BooLigero: Improved Sublinear Zero Knowledge Proofs for Boolean Circuits
Yaron Gvili, Sarah Scheffler, Mayank Varia
Financial Crypto 2021

Protecting Cryptography against Compelled Self-Incrimination
Sarah Scheffler, Mayank Varia
USENIX Security 2021

Arithmetic Expression Construction
Leo Alcock, Sualeh Asif, Jeffrey Bosboom, Josh Brunner, Charlotte Chen, Erik D. Demaine, Rogers Epstein, Adam Hesterberg, Lior Hirschfeld, William Hu, Jayson Lynch, Sarah Scheffler, Lillian Zhang
ISAAC 2020

Case Study: Disclosure of Indirect Device Fingerprinting in Privacy Policies
Julissa Milligan, Sarah Scheffler, Andrew Sellars, Trishita Tiwari, Ari Trachtenberg, Mayank Varia
STAST 2019

PSPACE-completeness of Pulling Blocks to Reach a Goal
Joshua Ani, Sualeh Asif, Erik D. Demaine, Yevhenii Diomidov, Dylan Hendrickson, Jayson Lynch, Sarah Scheffler, Adam Suhl
JCDCG^3 2019 / JIP 2020

From Soft Classifiers to Hard Decisions: How fair can we be?
Ran Canetti, Aloni Cohen, Nishanth Dikkala, Govind Ramnarayan, Sarah Scheffler, Adam Smith
ACM FAT* 2019

The Unintended Consequences of Email Spam Prevention
Sarah Scheffler, Sean Smith, Yossi Gilad, Sharon Goldberg
PAM 2018

Current Research

Transparency and Public Verifiability for Content Moderation in End-to-End Encryption

It is possible to build systems that analyze the content of end-to-end encrypted messaging, either by observing the content on the client side, or by obliviously processing it on the server side. However, these systems bring both technical and policy challenges for ensuring that these systems are used as advertised and without enabling mass surveillance or infringing on civil rights. Technical improvements are possible for a limited number of these issues, and the rest must be addressed at a legal or policy level.

Enabling Journalistic, Social, and Scientific Research with Applied Privacy-Preserving Computing

The last decade has seen an explosion in cryptographic and statistical tools for privacy-preserving computation: multi-party computation, differential privacy, homomorphic encryption, and (sort of) functional encryption. But these tools remain largey confined to theoretical proposals, or, at best, unwieldy codebases with high learning curves. I aim to bring this privacy-preserving computation to particular users who need high levels of privacy in their analysis, especially journalists wishing to protect sources, researchers needing to comply with privacy protocols. Actual journalists will be using the privacy-preserving functionality soon, via the Digital Witness Lab!

Choosing the right privacy-enhancing tool for the privacy-enhancing job

We have finally started to see some adoption of privacy-enhancing technologies like MPC, differential privacy, and homomorphic encryption at scale. But the deployment of these technologies does not always mean that the privacy problem is solved. What privacy gaps remain? How do we see these systems developing into the future?

Formalization of legal copyright notions

I have one minor contribution in the space of understanding U.S. Copyright's notion of "substantial similarity" a little better.