Sanjam Garg

Sanjam Garg

Auto-generated profile, based on publicly available information

Sanjam Garg

University of California, Berkeley, Berkeley, CA, USA

Research Focus & Metrics

Academic Career Length

15Years (2012-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
31 March 2018
Verified
Book Chapter
Closed access
PDF file
SGSanjam Garg
Sanjam Garg
ASAkshayaram Srinivasan
Akshayaram Srinivasan

4 November 2018
Verified
Book Chapter
Closed access
PDF file
SGSanjam Garg
Sanjam Garg
MHMohammad Hajiabadi
Mohammad Hajiabadi
MMMohammad Mahmoody
Mohammad Mahmoody
ARAhmadreza Rahimi
Ahmadreza Rahimi


24 December 2015
Verified
Book Chapter
Closed access
PDF file
SGSanjam Garg
Sanjam Garg
CGCraig Gentry
Craig Gentry
SHShai Halevi
Shai Halevi
MZMark Zhandry
Mark Zhandry

29 July 2017
Verified
Book Chapter
Closed access
PDF file
CCChongwon Cho
Chongwon Cho
NDNico Döttling
Nico Döttling
SGSanjam Garg
Sanjam Garg
DGDivya Gupta
Divya Gupta
PMPeihan Miao
Peihan Miao
APAntigoni Polychroniadou
Antigoni Polychroniadou

13 January 2025 in IACR Communications in Cryptology
Verified
Journal Article
Open access
PDF file
JFJaiden Fairoze
Jaiden Fairoze
SGSanjam Garg
Sanjam Garg
SJSomesh Jha
Somesh Jha
SMSaeed Mahloujifar
Saeed Mahloujifar
MMMohammad Mahmoody
Mohammad Mahmoody
MWMingyuan Wang
Mingyuan Wang
We present a publicly-detectable watermarking scheme for LMs: the detection algorithm contains no secret information, and it is executable by anyone. We embed a publicly-verifiable cryptographic signature into LM output using rejection sampling and prove that this produces unforgeable and distortion-free (i.e., undetectable without access to the public key) text output. We make use of error-correction to overcome periods of low entropy, a barrier for all prior watermarking schemes. We implement our scheme and find that our formal claims are met in practice.