Research map / Learning theory and algorithms / Sample complexity
When are Local Queries Useful for Robust Learning?
Pascale Gourdeau, Varun Kanade, Marta Kwiatkowska, James Worrell
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 changed | Where | Measured | Effect |
|---|---|---|---|
| Add ρ-local equivalence queries to random examples | Distribution-free exact-in-the-ball robust learning · Concept classes with robustly consistent learners | robust learnability | Improved |
| Use local equivalence queries with radius λ strictly smaller than ρ | Monotone conjunctions and superclasses · Distribution-free exact-in-the-ball robust learning | expected robust risk: \frac{1}{2} | Hurt |
| Replace local membership queries with local equivalence queries | Conjunctions · Distribution-free exact-in-the-ball robust learning | query complexity: O\left( \frac{1}{\epsilon}\left(n+\log\frac{1}{\delta}\right)\right) versus 2^{\rho-1} | Improved |
| Set the local equivalence query radius equal to the adversary's budget | Monotone conjunctions · Distribution-free robust learning | robust learnability | Improved |
| Set the local equivalence query radius below the adversary's budget | Monotone conjunctions and classes containing them · Distribution-free robust learning | robust learnability | Hurt |
| Restrict the adversary to precision-bounded counterexamples | Halfspaces in ℝ^n · Distribution-free ρ-robust learning | query complexity: m(n,\epsilon,\delta)\cdot\frac{B^2}{\tau^2} | Improved |
| Assume a margin only on the data distribution's support | Halfspaces in ℝ^n · Robust learning with ρ-LEQ | query complexity | Hurt |
| Require the margin condition throughout the ρ-expansion of the distribution support | Halfspaces in ℝ^n · Robust learning with ρ-LEQ | query complexity: r=\frac{mB^2}{\gamma^2} | Improved |
Closest papers by meaning
- Incorporating Unlabeled Data into Distributionally Robust Learning (JMLR 2021 · similarity 0.86)
- On the Hardness of Robust Classification (NeurIPS 2019 · similarity 0.86)
- Tradeoffs between Mistakes and ERM Oracle Calls in Online and Transductive Online Learning (NeurIPS 2025 · similarity 0.85)
- Regret–Oracle Complexity Tradeoffs in Agnostic Online Learning (NeurIPS 2026 · similarity 0.84)
- Sample and Computationally Efficient Robust Learning of Gaussian Single-Index Models (NeurIPS 2024 · similarity 0.84)
- Robust Regression of General ReLUs with Queries (NeurIPS 2025 · similarity 0.83)
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.
