New paper available soon, by Sarah Radway (Tufts PhD student), Miranda Christ (Columbia PhD student), and me: "Differential Privacy and Swapping: Examining De-Identification's Impact on Minority Representation and Privacy Preservation in the U.S. Census" x.com/binarybits/sta…

Video teaser by Sarah Radway and Miranda Christ for their paper "Differential Privacy and Swapping: Examining De-Identification's Impact on Minority Representation and Privacy Preservation in the U.S. Census" at @ieeessp: youtube.com/watch?v=YjFpAI…

it was wonderful presenting at National Academies this week with Miranda Christ — building upon our paper with Steven M. Bellovin, our work contextualizes the census bureau’s transition to dp, through an in-depth comparison with swapping for the single-year-of-age use case :)

Field notes from the recent Theory of Cryptography Conference put on by IACR via a16z crypto research intern Miranda Christ. 🙌 a16zcrypto.com/theory-of-cryp…

#ePrint Undetectable Watermarks for Language Models: M Christ, S Gunn, O Zamir ia.cr/2023/763

Enjoyed exploring zero knowledge range proofs during my internship MystenLabs.sui :)

Writing this paper with Joseph Bonneau during my a16z crypto internship has been one of the highlights of my PhD! Here it is in a more digestible blog post format. Read on to learn about the limits of stateless blockchains:

1/ New paper “Undetectable Selfish Mining” by Matt Weinberg and me: arxiv.org/abs/2309.06847 TLDR; 👇🧵

I've been playing this game called squaredle and it's really fun! They also made really cool merch 😄 If you also play, add me to your leader-board (I'm not very good!): squaredle.app/?i=MTUyMjExNw__

wish I could tell my younger self this too

ia.cr/2024/1711 New paper with Yuval Efron Miranda Christ and Joseph Bonneau. Delay functions let a group generate "perfect" randomness, even if only one participant is honest. In this work, we show that DFs are, in fact, necessary to get such randomness!

We had to wait but good things and our paper got into Eurocrypt 🇪🇸 eprint.iacr.org/2024/1711 We show that Delay are not just useful but necessary to construct dishonest-majority secure randomness beacons. Time to adopt VDFs CC: Yuval Efron Miranda Christ Joseph Bonneau

🤯 AI images getting too real? (Think GPT-4o, Imagen-3, etc!) Worried about trust in the GenAI era? We tackled this head-on at #ICLR2025! Presenting the first UNDETECTABLE watermark for generative models! 🚀 A big step forward w/ Sam Gunn & Dawn Song 🧵👇 (1/13)