Ph.D. student · Computer Science · Yale
Yunhao Wang
I am a fourth-year Ph.D. student at Yale University, advised by Katerina Sotiraki and Fan Zhang, and affiliated with IC3. Before Yale, I received my M.S. in Computer Science from Columbia University, working with Tal Malkin and Eran Tromer, and my B.S. in Computer Science and Applied Mathematics from UCLA.
New Encore, constant-size proofs for FHE computation, to appear at IEEE S&P 2027
Research
My research is in applied cryptography, centered on lattices. I design post-quantum protocols that are both provably secure and practical, and so far have worked on fully homomorphic encryption, metadata-private messaging, leader election for blockchains, and verifiable computation over encrypted data. Going forward, I am excited about advanced public-key encryption, new lattice-based primitives, and post-quantum cryptography and SNARKs more broadly, toward a future where strong cryptography is quantum-safe, expressive, and verifiable by default.
§01Publications Google Scholar ↗
Authors are listed in alphabetical order, unless marked with †.
2026
-
Encore: Aggregatable FHE-over-SNARK with Constant Proof Size
IEEE Symposium on Security and Privacy (S&P), 2027
tl;dr Lifts LPCP-based SNARKs into the FHE-over-SNARK paradigm: constant-size proofs for homomorphic computation with native SIMD, plus d-Encore, a distributed prover whose sub-proofs aggregate into one succinct proof.
-
UnifOMR: Oblivious Message Retrieval with Near-optimal Concrete Efficiency
ACM Conference on Computer and Communications Security (CCS), 2026
tl;dr Shows that OMR is at least as hard as PIR, then closes most of the gap: 20–1080× faster server runtime than the prior state of the art.
-
Relect: Single Secret Leader Election via FHE with Reduced Computation and Communication and Transparent Setup
ACM Conference on Computer and Communications Security (CCS), 2026
tl;dr An RLWE-based SSLE that drops Qelect's trusted setup while running up to 48× faster with up to 2× less communication.
-
Dinocchio: Distributed Prover for Ring Arithmetic
Cryptology ePrint Archive, 2026
tl;dr The first distributed SNARK for rings with constant proof size and verification time: roughly m-fold prover speedup with m sub-provers, scaling to circuits with ~2³² constraints.
2025
-
Lattice-based Multi-message Multi-recipient KEM/PKE with Malicious Security
International Conference on the Theory and Application of Cryptology and Information Security (ASIACRYPT), 2025
tl;dr Breaks existing lattice mmPKE schemes with maliciously crafted public keys, then builds the first lattice mmKEM that stays private against them, with ~8× smaller ciphertexts than Kyber for 100 recipients.
-
IND-CPA-D of Relaxed Functional Bootstrapping: A New Attack, A General Fix, and A Stronger Model
ACM Conference on Computer and Communications Security (CCS), 2025
tl;dr Recovers the secret key of every existing relaxed functional bootstrapping construction within minutes, even at negligible failure probability, then gives a general fix and a stronger security model.
-
Qelect: Lattice-based Single Secret Leader Election Made Practical
USENIX Security Symposium, 2025
tl;dr The first practical constant-round post-quantum SSLE, built from an RLWE-based multi-party randomizable commitment and SIMD threshold FHE.
-
Snake-eye Resistant PKE from LWE for Oblivious Message Retrieval and Robust Encryption
International Conference on the Theory and Applications of Cryptographic Techniques (EUROCRYPT), 2025
tl;dr Proves OMRp2 DoS-resistant under standard LWE, introduces snake-eye resistant PKE, and derives the first robust lattice PKE without KEM-DEM, yielding a DoS-resistant OMR 12× faster than OMRp2.
2024
-
Relaxed Functional Bootstrapping: A New Perspective on BGV/BFV Bootstrapping
International Conference on the Theory and Application of Cryptology and Information Security (ASIACRYPT), 2024
tl;dr Relaxes what bootstrapping must guarantee and folds in a free function evaluation, making BGV/BFV bootstrapping 1–2 orders of magnitude faster.
-
PerfOMR: Oblivious Message Retrieval with Reduced Communication and Computation
USENIX Security Symposium, 2024
tl;dr Ring-LWE clues make OMR far cheaper: a retrieval circuit evaluated 13.8× faster, or public keys about 235× smaller.
-
Group Oblivious Message Retrieval
IEEE Symposium on Security and Privacy (S&P), 2024
tl;dr OMR for group messaging: server cost grows very slowly with group size, and each recipient's cost is independent of it.
2023
-
Amortized Functional Bootstrapping in less than 7 ms, with Õ(1) Polynomial Multiplications
International Conference on the Theory and Application of Cryptology and Information Security (ASIACRYPT), 2023
tl;dr Batched bootstrapping of LWE ciphertexts that is efficient in practice, not just asymptotically: under 5 ms per binary gate and about 6.7 ms per arbitrary 9-bit function.
§02Academic Service
-
Program committee
- USENIX Security2027
-
Sub-reviewer
- CRYPTO2023–2024, 2026
- EUROCRYPT2025–2026
- ASIACRYPT2024
- TCC2025–2026
- PKC2026
- IEEE S&P2025
- ACM CCS2025–2026
- USENIX Security2024
- FOCS2026
- SODA2027
§03Background
-
2023 – now
Ph.D. in Computer Science
Yale University · advised by Katerina Sotiraki and Fan Zhang
-
2021 – 2023
M.S. in Computer Science
Columbia University · with Tal Malkin and Eran Tromer
-
2014 – 2018
B.S. in Computer Science and in Applied Mathematics
University of California, Los Angeles · cum laude