Fast and Private Submodular and $k$-Submodular Functions Maximization with Matroid Constraints

Authors: Akbar Rafiey, Yuichi Yoshida

ICML 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical In this paper, we study the problem of maximizing monotone submodular functions subject to matroid constraints in the framework of differential privacy. We provide (1 - 1/e)-approximation algorithm which improves upon the previous results in terms of approximation guarantee. This is done with an almost cubic number of function evaluations in our algorithm. Moreover, we study k-submodularity, a natural generalization of submodularity. We give the first 1/2-approximation algorithm that preserves differential privacy for maximizing monotone k-submodular functions subject to matroid constraints. The approximation ratio is asymptotically tight and is obtained with an almost linear number of function evaluations.
Researcher Affiliation Academia 1Department of Computing Science, Simon Fraser University, Burnaby, Canada 2National Institute of Informatics, Tokyo, Japan. Correspondence to: Akbar Rafiey <arafiey@sfu.ca>, Yuichi Yoshida <yyoshida@nii.ac.jp>.
Pseudocode Yes Algorithm 1 Differentially Private Continuous Greedy; Algorithm 2 Improved Differentially Private Continuous Greedy Algorithm; Algorithm 3 Differentially private k-submodular maximization with a matroid constraint; Algorithm 4 Improved differentially private k-submodular maximization with a matroid constraint
Open Source Code No The paper does not provide any statement about releasing source code or a link to a code repository.
Open Datasets No The paper refers to 'sensitive dataset D' as an input to the functions being maximized, but it does not specify or provide access information for any concrete, publicly available dataset used for training or evaluation. The examples like 'medical data, web search query data, salary data, social networks' are general types of data, not specific datasets with access details.
Dataset Splits No The paper is theoretical and does not conduct empirical experiments, therefore it does not provide information about validation splits.
Hardware Specification No The paper does not mention any specific hardware used for running experiments. It focuses on theoretical algorithms and their complexity.
Software Dependencies No The paper does not provide specific software names with version numbers (e.g., programming languages, libraries, or solvers) required to replicate the experiments or implement the algorithms.
Experiment Setup No The paper describes algorithms theoretically and does not include an experimental setup section with concrete hyperparameters, training configurations, or system-level settings.