Attacks on Ring-LWE
summary
Paper shows practical attacks on decision RLWE instances that come from PLWE-style errors when the defining polynomial has a root mod q or a small-order root; it uses the change-of-basis spectral norm to predict how the coefficient-wise Gaussian error is distorted and to select vulnerable fields. Implementations in Sage recover secrets for TLS-like parameters (example run finishes in 13.5 hours with ~20 samples) and break smaller parameter sets in seconds; attack loops over guesses mod q so cost scales roughly O(q).
AI-written summary · about
chapters (16)
- 0:00Introduction
- 0:38Homomorphic Encryption
- 2:35LWE and Ring LWE
- 4:43Hardness Assumptions
- 6:06Embeddings and PLWE
- 8:30Error Distribution
- 10:39Attack Overview
- 13:08Attack Mechanisms
- 15:13Required Properties
- 17:56Attack Extensions
- 20:55Weak RLWE Fields
- 23:26Construction Theorem
- 25:10Parameter Selection
- 28:56Experimental Results
- 31:00Number Theory Problems
- 32:47Q&A
related talks
01h 28mThe Learning With Errors Problem and Cryptographic Applications
01h 15mAlgebraic Lattices and Ring Learning with Errors
30mOn the Concrete Security of LWE with Small Secret
01h 04mThe Mathematics of Lattices II
31mIdentity-Based Encryption
56mContinuous LWE is as Hard as LWE
39mLattice Based Cryptography Primer
21mQuantum Proofs of Deletion for Learning with Errors
