Weakly Learning to Match Experts in Online Community

Authors: Yujie Qian, Jie Tang, Kan Wu

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

Reproducibility Variable Result LLM Response
Research Type Experimental We use two different genres of datasets: QA-Expert and Paper Reviewer, to validate the proposed model. Our experimental results show that the proposed model can significantly outperform (+1.5-10.7% by MAP) the state-of-the-art algorithms for matching users (experts) with community questions. We have also developed an online system to further demonstrate the advantages of the proposed method.
Researcher Affiliation Academia Yujie Qian , Jie Tang and Kan Wu Tsinghua University Massachusetts Institute of Technology yujieq@csail.mit.edu, jietang@tsinghua.edu.cn, wu-k14@mails.tsinghua.edu.cn
Pseudocode No The paper contains mathematical formulations and descriptions of algorithms but no explicitly labeled pseudocode or algorithm blocks.
Open Source Code Yes The code for this paper is publicly available.3 https://www.aminer.cn/match expert
Open Datasets Yes QA-Expert: this dataset is from an online international QA challenge2. It consists of 87,988 users and 47,656 questions with 43,913 invitations and 9,561 responses. ... 2https://biendata.com/competition/bytecup2016/
Dataset Splits No The paper states: "In each dataset, we randomly select 60% as the training data and the remaining as test data." It does not explicitly mention a separate validation dataset or its split.
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., CPU, GPU models, memory) used for running the experiments.
Software Dependencies No The paper mentions software like Word2Vec and SVM-Light, but it does not provide specific version numbers for these or any other key software dependencies required for reproducibility.
Experiment Setup No The paper mentions some empirical settings for baselines (e.g., "set the number of topics to 30" for LDA, "k = 2, b = 0.75" for BM25) and states that the learning algorithm uses stochastic gradient descent and takes "1,000-2,000 iterations to converge". However, it lacks specific hyperparameters for the main Weak FG model, such as learning rates, batch sizes, or detailed optimizer configurations for reproducibility.