ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2207.14030
15
15

Hardness of Agnostically Learning Halfspaces from Worst-Case Lattice Problems

28 July 2022
Stefan Tiegel
ArXivPDFHTML
Abstract

We show hardness of improperly learning halfspaces in the agnostic model, both in the distribution-independent as well as the distribution-specific setting, based on the assumption that worst-case lattice problems, such as GapSVP or SIVP, are hard. In particular, we show that under this assumption there is no efficient algorithm that outputs any binary hypothesis, not necessarily a halfspace, achieving misclassfication error better than 12−γ\frac 1 2 - \gamma21​−γ even if the optimal misclassification error is as small is as small as δ\deltaδ. Here, γ\gammaγ can be smaller than the inverse of any polynomial in the dimension and δ\deltaδ as small as exp(−Ω(log⁡1−c(d)))exp(-\Omega(\log^{1-c}(d)))exp(−Ω(log1−c(d))), where 0<c<10 < c < 10<c<1 is an arbitrary constant and ddd is the dimension. For the distribution-specific setting, we show that if the marginal distribution is standard Gaussian, for any β>0\beta > 0β>0 learning halfspaces up to error OPTLTF+ϵOPT_{LTF} + \epsilonOPTLTF​+ϵ takes time at least dΩ~(1/ϵ2−β)d^{\tilde{\Omega}(1/\epsilon^{2-\beta})}dΩ~(1/ϵ2−β) under the same hardness assumptions. Similarly, we show that learning degree-ℓ\ellℓ polynomial threshold functions up to error OPTPTFℓ+ϵOPT_{{PTF}_\ell} + \epsilonOPTPTFℓ​​+ϵ takes time at least dΩ~(ℓ2−β/ϵ2−β)d^{\tilde{\Omega}(\ell^{2-\beta}/\epsilon^{2-\beta})}dΩ~(ℓ2−β/ϵ2−β). OPTLTFOPT_{LTF}OPTLTF​ and OPTPTFℓOPT_{{PTF}_\ell}OPTPTFℓ​​ denote the best error achievable by any halfspace or polynomial threshold function, respectively. Our lower bounds qualitively match algorithmic guarantees and (nearly) recover known lower bounds based on non-worst-case assumptions. Previously, such hardness results [Daniely16, DKPZ21] were based on average-case complexity assumptions or restricted to the statistical query model. Our work gives the first hardness results basing these fundamental learning problems on worst-case complexity assumptions. It is inspired by a sequence of recent works showing hardness of learning well-separated Gaussian mixtures based on worst-case lattice problems.

View on arXiv
Comments on this paper