Improving Online Algorithms via ML Predictions

Authors: Manish Purohit, Zoya Svitkina, Ravi Kumar

NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental 4 Experimental results
Researcher Affiliation Collaboration Ravi Kumar Google Mountain View, CA ravi.k53@gmail.com Manish Purohit Google Mountain View, CA mpurohit@google.com Zoya Svitkina Google Mountain View, CA zoya@cs.cornell.edu
Pseudocode Yes Algorithm 1: A simple 1-consistent algorithm
Open Source Code No The paper does not provide an explicit statement about releasing source code or a link to a code repository for the methodology described.
Open Datasets No For all experiments, we set the cost of buying to b = 100 and the actual number of skiing days x is a uniformly drawn integer from [1, 4b]. The predicted number of days y is simulated as y = x + ϵ where ϵ is drawn from a normal distribution with mean 0 and standard deviation σ. We generate a synthetic dataset with 50 jobs where the processing time of each job is sampled independently from a Pareto distribution with an exponent of α = 1.1. The paper describes how the data was generated or simulated, but does not provide concrete access information (link, DOI, specific citation) for a publicly available dataset.
Dataset Splits No The paper describes experiments run over '10000 independent trials' and '1000 independent trials' and varying a parameter 'σ', which indicates repeated simulations for evaluation. However, it does not specify explicit training/validation/test dataset splits, as the focus is on the performance of online algorithms with given predictions rather than training a machine learning model.
Hardware Specification No The paper does not provide specific details about the hardware used to run the experiments, such as GPU/CPU models, processor types, or memory specifications.
Software Dependencies No The paper does not provide specific software dependencies with version numbers used for the experiments.
Experiment Setup Yes For all experiments, we set the cost of buying to b = 100 and the actual number of skiing days x is a uniformly drawn integer from [1, 4b]. The predicted number of days y is simulated as y = x + ϵ where ϵ is drawn from a normal distribution with mean 0 and standard deviation σ. We set λ = 0.5 for the deterministic algorithm... We set λ = ln(3/2) for the randomized algorithm. ... we set the predicted job length yi = xi + ϵi, where ϵi is drawn from a normal distribution with mean zero and standard deviation σ.