When are Local Queries Useful for Robust Learning?

Authors: Pascale Gourdeau, Varun Kanade, Marta Kwiatkowska, James Worrell

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical In this paper, we study learning models where the learner is given more power through the use of local queries, and give the first distribution-free algorithms that perform robust empirical risk minimization (ERM) for this notion of robustness. [...] We then bound the query complexity of these algorithms based on online learning guarantees and further improve these bounds for the special case of conjunctions. We finish by giving robust learning algorithms for halfspaces with margins on both {0, 1}n and Rn.
Researcher Affiliation Academia Pascale Gourdeau University of Oxford pascale.gourdeau@cs.ox.ac.uk Varun Kanade University of Oxford varunk@cs.ox.ac.uk Marta Kwiatkowska University of Oxford marta.kwiatkowska@cs.ox.ac.uk James Worrell University of Oxford james.worrell@cs.ox.ac.uk
Pseudocode No No structured pseudocode or algorithm blocks were found in the paper.
Open Source Code No Did you include the code, data, and instructions needed to reproduce the main experimental results (either in the supplemental material or as a URL)? [N/A]
Open Datasets No The paper is theoretical and does not conduct experiments on specific datasets. It discusses concept classes and distributions abstractly, but does not provide information on publicly available datasets for empirical training.
Dataset Splits No No specific dataset split information (exact percentages, sample counts, or detailed splitting methodology) is provided as the paper is theoretical and does not conduct experiments on specific datasets.
Hardware Specification No No specific hardware details (exact GPU/CPU models, processor types, or memory amounts) used for running experiments are mentioned, as the paper is theoretical and does not involve empirical experiments.
Software Dependencies No No specific ancillary software details (e.g., library or solver names with version numbers) are mentioned, as the paper is theoretical and does not involve empirical experiments.
Experiment Setup No No specific experimental setup details (concrete hyperparameter values, training configurations, or system-level settings) are provided, as the paper is theoretical and does not involve empirical experiments.