Online Algorithms with Uncertainty-Quantified Predictions

Authors: Bo Sun, Jerry Huang, Nicolas Christianson, Mohammad Hajiesmaili, Adam Wierman, Raouf Boutaba

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

Reproducibility Variable Result LLM Response
Research Type Experimental Figure 1 compares the empirical competitive ratios (CRs) of our proposed online algorithms in the setting of a multiple-instance ski rental problem. The setup details can be found in Appendix D.6.
Researcher Affiliation Academia 1David R. Cheriton School of Computer Science, University of Waterloo, Waterloo, Canada. 2Computing + Mathematical Sciences Department, California Institute of Technology, Pasadena, USA. 3Manning College of Information and Computer Sciences, University of Massachusetts Amherst, Amherst, USA.
Pseudocode Yes Algorithm 1 DSR: Deterministic algorithm for ski rental; Algorithm 2 RSR(y): Randomized algorithm for ski rental; Algorithm 3 PFA(G): Protection-function-based algorithm; Algorithm 4 Online deterministic algorithm with PIP for ski rental; Algorithm 5 Online learning algorithm with uncertainty-quantified predictions; Algorithm 6 Online learning algorithm for multiple-instance ski rental
Open Source Code No The paper does not contain any statement or link indicating that the source code for the described methodology is publicly available.
Open Datasets No The paper describes generating synthetic data for its experiments: "We generate T = 3000 instances, each with true skiing days nt sampled uniformly at random from {1, . . . , 8}." It does not refer to a publicly available or open dataset with access information.
Dataset Splits No The paper describes the generation of instances for its online algorithm evaluations but does not specify traditional training, validation, or test splits for a machine learning model.
Hardware Specification No The paper does not provide any specific details about the hardware used to run the experiments.
Software Dependencies No The paper does not list any specific software components with version numbers.
Experiment Setup Yes Setup. We set buying cost to B = 2. We generate T = 3000 instances, each with true skiing days nt sampled uniformly at random from {1, . . . , 8}. Synthetic PIP predictions are generated by sampling a point pt from a normal distribution N(nt, σ2 t ) and then taking the 90% confidence interval (ℓt, ut) = (pt z0.95σt, pt + z0.95σt). Here, σt is set to simulate oscillating good and bad predictors: the first 10 instances have σt = 0, followed by the next 10 with σt = 6, and repeating.