Two-timescale Derivative Free Optimization for Performative Prediction with Markovian Data

Authors: Haitong Liu, Qiang Li, Hoi To Wai

ICML 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Numerical experiments verify our analysis.
Researcher Affiliation Academia 1Department of Computer Science, ETH Zurich. The research was primarily conducted as a student at The Chinese University of Hong Kong. 2Department of System Engineering and Engineering Management, The Chinese University of Hong Kong, Hong Kong, China.
Pseudocode Yes Algorithm 1 DFO (λ) Algorithm
Open Source Code No The paper does not provide any statements about releasing open-source code or links to a code repository.
Open Datasets No The paper describes generating synthetic data for its experiments (e.g., 'auto-regressive (AR) process' for Quartic Loss, 'Markovian Pricing Problem', 'Markovian Performative Regression') rather than using publicly available datasets with specified access information.
Dataset Splits No The paper does not explicitly specify training, validation, or test dataset splits. It discusses results based on 'number of samples i' or total samples observed during the experiments.
Hardware Specification Yes All experiments are conducted on a server with an Intel Xeon 6318 CPU
Software Dependencies Yes using Python 3.7.
Experiment Setup Yes Unless otherwise specified, we use the step size choices in (8) for DFO (λ).