Neural Link Prediction over Aligned Networks

Authors: Xuezhi Cao, Haokun Chen, Xuejian Wang, Weinan Zhang, Yong Yu

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

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments demonstrate that MNN outperforms the state-of-the-art methods and achieves 3% to 5% relative improvement of AUC score across different settings, particularly over 8% for cold start scenarios.
Researcher Affiliation Academia Xuezhi Cao, Haokun Chen, Xuejian Wang, Weinan Zhang, Yong Yu APEX Data & Knowledge Management Lab Shanghai Jiao Tong University cxz,chenhaokun,xjwang,wnzhang,yyu@apex.sjtu.edu.cn
Pseudocode No The paper describes the network design with mathematical equations and text, but it does not include a clearly labeled pseudocode block or algorithm.
Open Source Code Yes The source code as well as the datasets are available online1. 1http://apex.sjtu.edu.cn/projects/34
Open Datasets Yes We conduct experiments using two sets of aligned social networks, provided by (Cao and Yu 2016a).
Dataset Splits No The paper mentions '80% links are used for training' but does not specify a validation split or percentage.
Hardware Specification No The paper does not provide specific details about the hardware used for experiments, such as CPU or GPU models, or memory specifications.
Software Dependencies No The paper does not mention specific software dependencies with version numbers (e.g., Python 3.x, TensorFlow 2.x).
Experiment Setup Yes For our multi-neural-network model (MNN), we set the embedding dimension k = 80, sampling rate α = 100, weighting parameter β = 0.5 and the regularization term γ = 0.1. We design each neural network to have 2 hidden layers between the product layer and output layer, with width of 100 and 50 respectively.