HodgeRank With Information Maximization for Crowdsourced Pairwise Ranking Aggregation

Authors: Qianqian Xu, Jiechao Xiong, Xi Chen, Qingming Huang, Yuan Yao

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

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments show that the proposed methods boost the sampling efficiency as compared to traditional sampling schemes and are thus valuable to practical crowdsourcing experiments.
Researcher Affiliation Collaboration Qianqian Xu,1,3 Jiechao Xiong,2,3 Xi Chen,4 Qingming Huang,5,6 Yuan Yao7,3, 1 SKLOIS, Institute of Information Engineering, CAS, Beijing, China, 2 Tencent AI Lab, Shenzhen, China 3 BICMR and School of Mathematical Sciences, Peking University, Beijing, China 4 Department of IOMS, Stern School of Business, New York University, USA 5 University of Chinese Academy of Sciences, Beijing, China, 6 IIP., ICT., CAS, Beijing, China 7, Department of Mathematics, Hong Kong University of Science and Technology, Hong Kong
Pseudocode Yes Algorithm 1: Unsupervised active sampling algorithm. Input: An initial graph Laplacian L0 defined on the graph of n nodes. and Algorithm 2: Online supervised active sampling algorithm for binary comparison data. Input: Prior distribution parameters γ, μ0, L 1 0,γ.
Open Source Code No The paper does not contain a clear statement about open-sourcing the code for the described methodology or a link to a code repository.
Open Datasets Yes The first example gives a comparison of these three sampling schemes on a complete & balanced video quality assessment (VQA) dataset (Xu et al. 2011). It contains 38,400 paired comparisons of the LIVE dataset (LIV 2008) from 209 random observers.
Dataset Splits No The paper describes generating simulated data and using existing real-world datasets, but it does not specify explicit train/validation/test dataset splits (e.g., 80/10/10 percentages or sample counts).
Hardware Specification Yes All computation is done using MATLAB R2014a, on a Mac Pro desktop PC, with 2.8 GHz Intel Core i7-4558u, and 16 GB memory.
Software Dependencies Yes All computation is done using MATLAB R2014a
Experiment Setup Yes A crucial question here is how to choose γ in supervised active sampling experiments. In practice, for dense graph, we find that γ makes little difference in the experimental results, a smaller γ, say 0.01, or even 1e 5 is sufficient. However, for sparse graph such as reading level dataset, a bigger γ (i.e., γ = 1) may produce better performance.