Asynchronous Distributed Bilevel Optimization

Authors: Yang Jiao, Kai Yang, Tiancheng Wu, Dongjin Song, Chengtao Jian

ICLR 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Thorough empirical studies on public datasets have been conducted to elucidate the effectiveness and efficiency of the proposed ADBO.
Researcher Affiliation Academia Yang Jiao Tongji University Kai Yang Tongji University Tiancheng Wu Tongji University Dongjin Song University of Connecticut Chengtao Jian Tongji University
Pseudocode Yes Algorithm 1 ADBO: Asynchronous Distributed Bilevel Optimization
Open Source Code Yes Codes are available in https://github.com/ICLR23Submission6251/adbo.
Open Datasets Yes In data hypercleaning task, experiments are carried out on MNIST (Le Cun et al., 1998) and Fashion MNIST (Xiao et al., 2017) datasets.
Dataset Splits No The paper mentions "training and validation datasets on ith worker" in the problem formulation but does not specify the actual split percentages, counts, or methodology used for these splits to reproduce the data partitioning.
Hardware Specification No The paper does not provide specific hardware details such as CPU models, GPU models, or memory configurations used for running the experiments. It generally refers to "workers" and a "master" in a distributed system.
Software Dependencies No The paper mentions using "SGD optimizer" and a "logistic regression model" but does not specify software versions for libraries, frameworks, or programming languages (e.g., Python, PyTorch, TensorFlow versions).
Experiment Setup Yes In MNIST and Fashion MNIST datasets, we set N = 18, S = 9 and τ = 15. According to Cohen et al. (2021), we assume that the communication delay of each worker obeys the heavy-tailed distribution. The step-sizes are summarized in Table 2.