Pratyay Mukherjee

Great program #Nordicrypt_2025

Abstract. A long line of work has investigated the design of scalable secure multiparty computation (MPC) protocols with computational and communication complexity independent of the number of parties (beyond any dependence on the circuit size). We present the first unconditionally-secure MPC protocols for arithmetic circuits over {} with total computation 𝒪(|C|log|F|), where |C| and |F| denote the circuit and field size, respectively. Prior work could either achieve similar complexity only in {}, or required highly structured circuits, or expensive circuit transformations. To obtain our results, we depart from the prior approach of share packing in linear secret-sharing schemes; instead, we use an “unpacking” approach via {} secret sharing.

Scalable Multiparty Computation from Non-linear Secret Sharing (Sanjam Garg, Abhishek Jain, Pratyay Mukherjee, Mingyuan Wang) ia.cr/2025/1007

New work on Verifiable Randomness

Overleaf is down! Quintessential single point of failure!

The AI fix feature on @overleaf.com is pretty cool actually, I am loving it so far :-)

(and it's not insanely expensive either)

Why Bhutan is on the red list? 😶 #travelban

Full version at eprint.iacr.org/2024/1713.pdf

We instantiate the HE with class-group encryption, for which we show the crucial additional properties hold.

n/n

In this work, we propose a technique by using homomorphic encryption (HE) with certain properties, which translates any sigma-ZKP for algebraic relation to support straight-line extraction, without repeating.

4/n

For UC security of the entire scheme, the underlying ZKP must have a straight-line extractability. Existing techniques to achieve straight-line extractability rely on repeating the proof multiple times.

3/n

Straightline extraction is a crucial property of ZKP, which is essential for UC security. E.g., an optimized threshold BLS signature may employ a non-interactive ZKP for partial verification (instead of pairing) to attest that Hash(msg) and g_2 have the same exponent (Chaum-Pederson proof).

2/n

New paper PKC'25!

Joint work with Megan Chen (BU), Pousali Dey (ISI), Chaya Ganesh (IISC), Pratik Sarkar (Supra) and Swagata Sasmal (ISI)

1/n

New paper on ePrint on Tracebale CCA-secure Encryptions!

Kolkata Crypto Day (took place on 8th Dec at ISI Kolkata) slides are up!

sites.google.com/d/1Lz8Mff3Uj...

New work on ePrint!

Yeah, I was wearing one ;)

I love this idea!

Submission deadline approaching for the IACR Communications in Cryptology (CiC).

Don’t miss your chance to contribute to this diamond open access journal.

Deadline: January 13, 2025 (AoE)

Details and submission guidelines: cic.iacr.org/callforpapers

#cryptography #iacr

Thank you! Also hope you enjoyed the boat ride :-) #Asiacrypt2024

Paul Kocher will deliver an IACR Distinguished Lecture tomorrow at #Asiacrypt2024

Yeah, it was a last minute change unfortunately :-( sorry for any inconvenience (though, I have seen people not following it strictly), hope it worked ok?

#Asiacrypt2024 Rump Session submission is open now...
asiacrypt.iacr.org/2024/rumpses...

Adi's place?

#Asiacrypt2024 (Important Change): The welcome reception will take place at the Calcutta Rowing Club (15, Dhakuria, Rabindra Sarobar, Kolkata, West Bengal 700029). Be aware that the venue *dress code* prohibits slippers/flip-flops/open chappals and shirts without collars

#Asiacrypt2024 tip: Unfortunately, the air quality here is pretty bad at this time of year so using a mask (better if N95) is highly recommended if you are outside to avoid allergies.

#Asiacrypt2024 tip: For commuting use your Uber app; I have tried with an international credit card, and it works fine. There are many other apps, but for most of them possibly you would need an Indian phone number.

#Asiacrypt2024 is almost here. As someone who was born and raised in Kolkata, I decided to provide some unofficial updates/tips and will post them as they come into my mind. Folks, especially who are coming from outside the sub-continent, might find these useful.

1/3

I don’t understand why the hotcrp for IACR conferences always need signing in plus the password autofill never works, unlike for security conferences (CCS, S&P, Usenix) 🥲

Happening with Asiacrypt