Looking to learn about lattice-based cryptography? Check out my tutorial and survey: https://t.co/DFH93Jx9pP

🎉 Congrats to Prof. @ChrisPeikert and co-authors on receiving the Distinguished Paper Award at CCS 2025 for groundbreaking work in secure threshold decryption for fully homomorphic encryption! 👏 🔗 https://t.co/6Vw1SzwcA9 #UMCSE #UMich #Cryptography #Cybersecurity #CCS2025

🚀 Michigan CSE is hiring! @UMichCSE Open tenure-track roles: • AI + Education (Assoc./Full Prof) • AI Foundations (Assoc./Full Prof) • An open-rank, tenure-track position in CSE 📅 Apply by Nov 30, 2025 🔗 https://t.co/g0jMS9pVgE #FacultySearch #CSE #AI @UMengineering

Study breaks and fall colors — a perfect combo 🍁

We think these tools will have many more benefits and applications in FHE and beyond, like homomorphic linear algebra, proof systems, etc. (And in any case, the underlying math is beautiful! 🤓) Check it out here:

Most notably, we get optimal plaintext packing for any finite-field slot type, and efficient packed bootstrapping that uses this parallelism. Key tools: structured (short & CRT) ring bases, and efficient CRT transforms—both "in the clear," and homomorphically via automorphisms.

This is our starting point. We adopt the very general setting of 𝙖𝙗𝙚𝙡𝙞𝙖𝙣 (Galois) number fields—which by Kronecker–Weber are equivalent to cyclotomic subfields—and give a broad collection of tools for doing lattice-based crypto and FHE in them.

In 2017, Arita & Handa proposed a neat alternative: use the 𝙙𝙚𝙘𝙤𝙢𝙥𝙤𝙨𝙞𝙩𝙞𝙤𝙣 𝙨𝙪𝙗𝙧𝙞𝙣𝙜 of a prime 𝘱. This gives an optimal number of 𝔽_𝘱-slots: the degree of the subring! But many parts of FHE weren't addressed, like bootstrapping and homom. linear algebra.

So, applications often must settle for a (much) larger slot type that extends the desired one, like 𝔽_{2^24} instead of 𝔽_{2^8}. This results in fewer slots (by 3x in this example), and "wasted potential" for SIMD parallelism. Can we do better?

~All FHE schemes use 𝙘𝙮𝙘𝙡𝙤𝙩𝙤𝙢𝙞𝙘 rings for efficiency. They have fast ring arithmetic, nice geometry, and homomorphic ops on plaintext vectors with many "slots." 𝘽𝙐𝙏: cyclotomics with the desired dimension and "slot type" (e.g., 𝔽₂₅₆) often 𝙙𝙤𝙣'𝙩 𝙚𝙭𝙞𝙨𝙩.

Thrilled to finally share this ⚡𝙣𝙚𝙬 𝙥𝙖𝙥𝙚𝙧⚡ with my (now-graduated!) student Zachary Pepin, which will appear at TCC 2025. We tackle a common inconvenience in BGV/BFV-style FHE: getting the desired kind of "SIMD slots" for plaintext packing. 🧵 https://t.co/MmUYV1r2i8

Huge congratulations to Prof. Michał Dereziński on receiving the 🏆 2025 @Google ML & Systems Junior Faculty Award for his work on theoretical ML and randomized algorithms! 🔽Click image for more https://t.co/YC3Vvf6fBI #MachineLearning #RandomizedAlgorithms #ResearchExcellence

🎉 Congrats to @eleanormlin and Dylan Zapzalka on receiving NSF Graduate Research Fellowships! A proud moment for Michigan CSE as these rising stars are recognized for their outstanding research potential. 👏 #NSFGRFP #UMichCSE #STEM #ComputerScience 🔽 https://t.co/GYrCk3lqcr

It’s always a good time to study lattices!

Great @tradeoffspod interview on the implementation challenges state Medicaid programs are staring down. "There have been exactly zero successes at implementing work requirements in this country… we've got two relatively well-documented struggle buses." https://t.co/15Wouwe6Xt

We’re happy to announce that @GabrielPoesia will be joining our faculty as an assistant professor in Fall 2026. Welcome to CSE! ▶️Learn more about Gabriel here: https://t.co/WD0dcIDWVR #UMichCSE #GoBlue

Ontario Crypto Day coming up: https://t.co/VnpyhuFkdU

THE MICHIGAN WOLVERINES ARE THE 2025 NATIONAL CHAMPIONS 🏆

Try to imagine a US President telling Ford that they shouldn't sell cars to foreigners, and that they should cut their R&D. It wouldn't happen. But that's what we're doing to one of America's most important exports: Education. 🧵

CSE PhD student @diwenxue has received a 🏆 Rackham Predoctoral Fellowship award recognizing his outstanding research accomplishments and his work on internet security and measurement. 👇Click below for more on Diwen's research [article]:

“ Efficient Proofs of Possession for Legacy Signatures” https://t.co/tpgy3R05se