Deletion Robust Submodular Maximization over Matroids
Authors: Paul Duetting, Federico Fusco, Silvio Lattanzi, Ashkan Norouzi-Fard, Morteza Zadimoghaddam
ICML 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In the experiments we evaluate the performance of our deletion robust centralized and streaming algorithms on real world data. |
| Researcher Affiliation | Collaboration | 1Google Research, Z urich, Switzerland 2Department of Computer, Control and Management Engineering Antonio Ruberti , Sapienza University of Rome, Rome, Italy. |
| Pseudocode | Yes | Algorithm 1 Centralized Algorithm Phase I... Algorithm 2 Algorithm Phase II... Algorithm 3 Streaming Algorithm Phase I... Algorithm 4 Lazy Greedy... Algorithm 5 Swapping algorithm |
| Open Source Code | No | The information is insufficient. The paper does not explicitly state that the code for their methodology is open-source or provide a link to it. A GitHub link is provided for a dataset used, not for the implementation of their algorithms. |
| Open Datasets | Yes | We use the Movie Lens 1M database (Harper & Konstan, 2016)... We use the Facebook dataset from Mc Auley & Leskovec (2012)... In our experiments, we use two datasets. Run In Rome (Fusco, 2022), that contains 8425 positions recorded by running activity in Rome, Italy and a random sample (10351 points) from the Uber pickups dataset (Kaggle, 2020). |
| Dataset Splits | No | The information is insufficient. The paper describes the datasets used and the experimental setup (e.g., number of runs, random permutations for streaming), but does not explicitly provide details about training, validation, or test dataset splits or a cross-validation setup. |
| Hardware Specification | No | All the experiments were run on a common computer, and running them on any other device would not affect the results in any way. |
| Software Dependencies | No | The information is insufficient. The paper describes algorithms used as subroutines but does not provide specific software dependencies or library version numbers required to reproduce the experiments. |
| Experiment Setup | Yes | For our experiments we set this precision parameter of lazy greedy ε0 to 0.0001. ... Finally, parameter α balances the trade off between the two terms; in our experiments it is set to 0.95. ... We set h to be the empirical standard deviation of the pairwise distances in the Run In Rome dataset, while we set h2 = 5000 in the Uber Dataset. ... α is a regularization parameter (that we set to 10 in the experiments). |