Cross-Lingual Taxonomy Alignment with Bilingual Biterm Topic Model

Authors: Tianxing Wu, Guilin Qi, Haofen Wang, Kang Xu, Xuan Cui

AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct experiments on two kinds of real world datasets. The experimental results show that our approach significantly outperforms the designed state-of-the-art comparison methods.
Researcher Affiliation Academia 1 Key Laboratory of Computer Network and Information Integration of State Education Ministry, School of Computer Science and Engineering, Southeast University, China {wutianxing, gqi, kxu, xcui}@seu.edu.cn 2 East China University of Science & Technology, China whfcarter@ecust.edu.cn
Pseudocode Yes Algorithm 1: Generative Process of Bi BTM
Open Source Code No The paper mentions a public GitHub link, but it specifies that the link is for 'two different kinds of real world datasets, which are publicly available2. 2https://github.com/jxls080511/080424'. There is no statement indicating that the source code for the methodology is available.
Open Datasets Yes We evaluated our proposed approach on two different kinds of real world datasets, which are publicly available2. 2https://github.com/jxls080511/080424
Dataset Splits No The paper describes how ground truth data was generated ('five annotators labelled the most relevant category... for each of the 100 randomly selected categories') and mentions '500 iterations of Gibbs sampling to converge' but does not specify train/validation/test dataset splits for model training or evaluation, nor does it refer to standard predefined splits.
Hardware Specification No The paper does not provide any specific hardware details such as GPU models, CPU types, or memory specifications used for running the experiments.
Software Dependencies No The paper mentions tools like 'Google Translate' and 'Fudan NLP (Qiu, Zhang, and Huang 2013)' but does not provide specific version numbers for these or any other software dependencies crucial for replication.
Experiment Setup Yes In Bi BTM, we set α = 50/K, β = 0.1 and K = 120 (the empirical tuning results will be presented in Section 4.2.2). In Bi LDA, we set α = 50/K, β = 0.1 and K = 80 (the empirical tuning results will be presented in Section 4.2.2). For each model, we ran 500 iterations of Gibbs sampling to converge.