Optimal Hessian/Jacobian-Free Nonconvex-PL Bilevel Optimization

Authors: Feihu Huang

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

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct some numerical experiments on the bilevel PL game and hyper-representation learning task to demonstrate efficiency of our proposed method.
Researcher Affiliation Academia 1College of Computer Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing, China 2MIIT Key Laboratory of Pattern Analysis and Machine Intelligence, Nanjing, China.
Pseudocode Yes Algorithm 1 Hessian/Jacobian-free Bilevel Optimization (i.e, HJFBi O) Algorithm
Open Source Code No No explicit statement or link to open-source code for the described methodology is provided in the paper.
Open Datasets No The paper describes generating synthetic data for its experiments ('samples {pi}n i=1, {qi}n i=1, {r1 i }n i=1 and {r2 i }n i=1 are independently drawn from normal distributions' and 'We randomly generate n = 30 d samples of sensing matrices {Ci}n i=1 from standard normal distribution, and then compute the corresponding no-noise labels oi = Ci, H '). It does not provide access information for a publicly available or open dataset.
Dataset Splits Yes We split all samples into two dataset: a train dataset Dt with 40% data and a validation dataset Dv with 60% data.
Hardware Specification No No specific hardware details (e.g., exact GPU/CPU models, memory amounts) used for running the experiments are provided in the paper.
Software Dependencies No No specific software dependencies with version numbers are provided in the paper.
Experiment Setup Yes For fair comparison, we set a basic learning rate as 0.01 for all algorithms. In our HJFBi O method, we set δϵ = 10 5.