Optimal Sampling Gaps for Adaptive Submodular Maximization

Authors: Shaojie Tang, Jing Yuan8450-8457

AAAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct experiments to evaluate the impact of probability sampling based on two popular machine learning applications: adaptive viral marketing and pool-based active learning [Golovin and Krause 2011]. ... We present the results in Figure 1 and Figure 2. We observe that the performance trends of the algorithms are overall consistent between all datasets and application domains.
Researcher Affiliation Academia Shaojie Tang,1 Jing Yuan, 2 1Naveen Jindal School of Management, University of Texas at Dallas 2 Department of Computer Science, University of Texas at Dallas shaojie.tang@utdallas.edu, csyuanjing@gmail.com
Pseudocode No The paper describes algorithms (e.g., adaptive greedy algorithm) but does not provide them in structured pseudocode or algorithm blocks.
Open Source Code No The paper does not contain an explicit statement about releasing open-source code for its methodology, nor does it provide a link to a code repository.
Open Datasets Yes We capture the social network by a directed weighted graph and run experiments on four large-scale benchmark social networks: Wikivote, Net HEPT, Net PHY and Epinions (http://snap.stanford.edu/data/).
Dataset Splits No The paper mentions training models and using datasets but does not explicitly specify training, validation, or test splits with percentages, counts, or references to standard split methodologies.
Hardware Specification No The paper mentions running experiments and simulations but does not specify any hardware details such as GPU models, CPU models, or memory.
Software Dependencies No The paper does not provide specific software dependencies with version numbers, such as Python versions or library versions.
Experiment Setup Yes We set k = 10 for Epinions, and k = 20 for other three datasets. ... We sample each node independently at a sampling rate r, that is, each node is being sampled independently with probability equal to r. We vary r from 0.1 to 1. ... For each sampling rate, we obtain 30 samples, and report the average performance of each algorithm over these samples...