Hyperbolic Heterogeneous Information Network Embedding

Authors: Xiao Wang, Yiding Zhang, Chuan Shi5337-5344

AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct extensive experiments to evaluate the performance of HHNE in terms of representation capacity and generalization ability on two real-world datasets. The results show the superiority of HHNE by comparing with the state-of-the-art techniques.
Researcher Affiliation Academia School of Computer Science, Beijing University of Posts and Telecommunications {xiaowang, zyd, shichuan}@bupt.edu.cn
Pseudocode No The paper describes the model and optimization steps mathematically but does not include a distinct section or figure labeled "Pseudocode" or "Algorithm".
Open Source Code No The paper does not provide any explicit statement about releasing the source code for the described methodology or a link to a code repository.
Open Datasets Yes DBLP is a bibliographic dataset in computer science. We use a subject of DBLP, i.e., DBLP-4-Area taken of (Sun et al. 2011)... Movie Lens1 contains knowledge about movies (Cantador, Brusilovsky, and Kuflik 2011). We extract a subset of from Movie Lens... 1https://grouplens.org/datasets/hetrec-2011/
Dataset Splits No We split the chosen edges and negative samples into validation and test. In our experiments, we train the embeddings on the residual network, and use the validation data to tune the model parameters. The paper mentions using validation data but does not provide specific split percentages, counts, or a detailed methodology for the validation split.
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., GPU/CPU models, memory) used to run the experiments.
Software Dependencies No The paper does not mention specific software components with version numbers (e.g., Python 3.8, PyTorch 1.9) required to reproduce the experiments.
Experiment Setup Yes For random walk based methods Deep Walk, node2vec, metapath2vec and HHNE, we set neighborhood size as 5, walk length as 80, walks per node as 40. For LINE, metapath2vec, Poincar e Emb and HHNE, we set the number of negative samples as 10.