A Regret Minimization Approach to Iterative Learning Control

Authors: Naman Agarwal, Elad Hazan, Anirudha Majumdar, Karan Singh

ICML 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We provide theoretical and empirical evidence that the proposed algorithm outperforms existing methods on several benchmarks. ... We experimentally demonstrate that the algorithm yields substantial improvements over ILC in linear and non-linear control settings. ... We demonstrate the efficacy of the proposed approach on two sets of experiments:
Researcher Affiliation Collaboration Naman Agarwal 1 Elad Hazan 1 2 Anirudha Majumdar 1 2 Karan Singh 3 1Google AI Princeton, Princeton, NJ, USA 2Princeton University, Princeton, NJ, USA 3Microsoft Research, Redmond, WA, USA.
Pseudocode Yes Algorithm 1 i GPC Algorithm ... Algorithm 2 GPCRollout ... Algorithm 3 Nested-OCO Algorithm
Open Source Code Yes The implementations along with some further experiments are present at https: //github.com/Min Regret/deluca-igpc.
Open Datasets No The paper describes using simulated environments (Double Integrator, Quadrotor with Wind, Reacher with Impulse) for its experiments, which are custom simulations or theoretical models rather than publicly available datasets with specified access information (link, DOI, or formal citation).
Dataset Splits No The paper does not provide specific details on dataset splits (e.g., percentages, sample counts for training, validation, or testing sets). The experiments are conducted on simulated environments, and the concept of standard data splits is not applied or described.
Hardware Specification No The paper does not specify the hardware used for running its experiments, such as specific GPU/CPU models, processors, or cloud computing resources.
Software Dependencies No The paper mentions 'JAX-based (Bradbury et al., 2018) differentiable implementations' but does not provide specific version numbers for JAX or any other software dependencies, which are necessary for full reproducibility.
Experiment Setup No The paper states: 'For further details on the setups and hyperparameter tuning please see Appendix (Section E).' This indicates that specific experimental setup details, including hyperparameters, are not provided in the main text.