AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

SALSA VERDE: a machine learning attack on LWE with sparse small secrets

Cathy Yuanchen Li, Emily Wenger, Zeyuan Allen-Zhu, Francois Charton, Kristin E. Lauter

NeurIPS 2023 · Adversarial robustness and security › Attacks on text and graphs · approach: Membership inference and inversion · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
reorder rows of the lattice matrix before reduction (ablation)VERDE preprocessing · LWE n=256, log2 q=20reduction factor: 0.53 -> 0.47Improved
reduce BKZ floating point precision (ablation)VERDE preprocessing · LWE n=256, log2 q=20preprocessing time (hrs/CPU): 136 -> 24Improved
lower omega from 15 to 10 (ablation)VERDE preprocessing · LWE n=256, log2 q=20reduction factor: 0.47 -> 0.43Mixed
interleave BKZ 2.0 with efficient reduction, adaptive blocksize, early stopping (ablation)VERDE preprocessing · LWE n=256, log2 q=20preprocessing time (hrs/CPU): 38 -> 7.5Improved
run distinguisher on held-out preprocessed vectors instead of random vectors (ablation)VERDE secret recovery · n=256, log2 q=23, binaryfraction recovered: 1/5 -> 2/5Improved
use smaller base B=5363 instead of 7150 (ablation)VERDE tokenization · n=256, log2 q=16, Gaussian h=6fraction recovered: 1/10 -> 3/10Improved
run 5 model initializations instead of 1 (ablation)VERDE · n=256, log2 q=12, binary h=4fraction recovered: 1/10 -> 3/10Improved
increase preprocessing time from 7 to 35 days (ablation)VERDE · n=512, log2 q=41, binaryhighest h recovered: 20 -> 63Improved
lower modulus from log2 q=20 to 12VERDE · n=256, binaryhighest h recovered: 33 -> 6Hurt
raise NoMod percentage of training data above thresholdVERDE · n=256, log2 q=20, binarysecret recovery: 67%Improved
permute columns of A and re-preprocessVERDE · n=256, log2 q=14, binary h=6-8secret recoveryImproved
replace seq2seq with encoder-only transformer with EMD lossVERDE model · n=256 and 512epochs to recoveryMixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy