search

fani-lab / OpeNTF

Neural machine learning methods for Team Formation problem.
Other
20 stars 15 forks source link

Research on heterogeneous GNNs #99

Open karan96 opened 2 years ago

karan96 commented 2 years ago

Research on how to generate the heterogeneous graph and use it to generate graph embeddings and use them as the encoding part and decoding part.

karan96 commented 2 years ago

There are few papers that I am adding here with their link related to GNNs that I have started reviewing, kindly give them a look and suggest if they are worth the time or not. Please feel free to add other resources in this issue that you may have found. Semi-supervised User Geolocation via Graph Convolutional Networks - https://arxiv.org/pdf/1804.08049.pdf Metapath-guided Heterogeneous Graph Neural Network for Intent Recommendation - https://dl.acm.org/doi/abs/10.1145/3292500.3330673?casa_token=PQl7qvGOLW4AAAAA:CMCaOtDJJg1P0qDy270sCOTBVxQTdsluzoZbcs63VkOhr491-frmGOA8qGHPoaKTk8J1Jwi_Vd0 Graph neural networks: A review of methods and applications - https://www.sciencedirect.com/science/article/pii/S2666651021000012?via%3Dihub

hosseinfani commented 2 years ago

These are good papers to give you an idea of using gnn. However, their main focus is not considering geolocation. The first work tries to predict the location of users (output) The second work does not consider location information, does it? The last one is a good survey on applications of gnn. But misses those that consider location, I guess.

How about this https://dl.acm.org/doi/abs/10.1145/3404835.3463105 . Although it does not have the location or similar word in the title or abstract, it considers the location (venue) of the papers, not the authors.

karan96 commented 2 years ago

The second paper I selected, as I thought it, will give me an idea about how meta-paths are calculated in a GNN. The paper could be a good approach. I will get to reading these.

karan96 commented 2 years ago

@hosseinfani Hi Dr. Fani, Could you please help me with some of the papers to go through for GNN that might include location, I am having a hard time finding the most relevant paper that most closely align with our research idea.

hosseinfani commented 2 years ago

@karan96 I would browse the papers in the proceeding of good conferences paper by paper (instead of searching). E.g, https://dblp.org/db/conf/www/www2021.html

1-(https://dblp.org/pid/50/7648.html): STAN: Spatio-Temporal Attention Network for Next Location Recommendation. 2177-2185