The paper establishes LWE-based near-optimal hardness for agnostic learning, one-sided reliable learning, and fairness auditing of homogeneous halfspaces under Gaussian marginals, extending prior results and narrowing the gap to upper bounds.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Near-Optimal Cryptographic Hardness of Learning With Homogeneous Halfspaces Under Gaussian Marginals
The paper establishes LWE-based near-optimal hardness for agnostic learning, one-sided reliable learning, and fairness auditing of homogeneous halfspaces under Gaussian marginals, extending prior results and narrowing the gap to upper bounds.