Triple Eagle: Simple, Fast and Practical Budget-Feasible Mechanisms

Authors: Kai Han, You Wu, He Huang, Shuang Cui

NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct extensive experiments to evaluate the empirical performance of our BFMs, and the experimental results strongly demonstrate the efficiency and effectiveness of our approach.
Researcher Affiliation Academia Kai Han School of Computer Science and Technology Soochow University, P.R.China hankai@suda.edu.cn You Wu School of Computer Science and Technology Soochow University, P.R.China 20235227129@stu.suda.edu.cn He Huang School of Computer Science and Technology Soochow University, P.R.China huangh@suda.edu.cn Shuang Cui School of Computer Science and Technology University of Science and Technology of China lakers@mail.ustc.edu.cn
Pseudocode Yes Algorithm 1: Triple Eagle Ran(α), Algorithm 2: LSDPricing(A, C, η), Algorithm 3: Triple Eagle Det(α), Algorithm 4: Triple Eagle Nm(α)
Open Source Code Yes The code of our paper can be found at: https://anonymous.4open.science/r/Triple Eagle-4D1B/README.md
Open Datasets Yes We use three real social network datasets: (1) Flixster [5] with 28,843 nodes and 272,786 edges; (2) Epinions [29] with 75,879 nodes and 508,837 edges; and (3) Slashdot [29] with 82,168 nodes 948,464 edges. [...] Following [8, 16, 20, 22], we use the CIFAR-10 dataset [8, 16, 20, 22] containing ten thousands 32 32 color images
Dataset Splits No The paper mentions using datasets like Flixster, Epinions, Slashdot, and CIFAR-10, but it does not provide specific train/validation/test splits, percentages, or explicit methodologies for splitting the data.
Hardware Specification Yes All the algorithms are implemented using C++ and are run on a Linux server with Intel Xeon Gold 6126 @ 2.60GHz CPU and 128GB memory.
Software Dependencies No The paper states "All the algorithms are implemented using C++" but does not provide specific version numbers for C++ or any other software dependencies, libraries, or frameworks used.
Experiment Setup Yes We adopt the well-known Independence Cascade (IC) model [27] for the influence spread function f( ), and follow [14] to set the activation probability pu,v of each edge (u, v) to 1/Nin(v), where Nin(v) denotes the set of in-neighbors of v. The cost c(u) of each node is generated uniformly at random from the interval [0, 1]. Each implemented randomized algorithm is independently executed 50 times, and the average result is reported.