naganandy / graph-based-deep-learning-literature Goto Github PK

Hi and thanks a lot for the cool repo. Very comprehensive. Wanted to advertise our NeurIPS paper from 2017. It’s an unsupervised GNN (the first as far as I can tell for node classification). Unfortunately similar to GraphSage which came out in the same conference :-). Here it is:

https://papers.nips.cc/paper/7097-learning-graph-representations-with-embedding-propagation.pdf

So these are just ideas for discussion:

1. Do You want to include papers from the Graph Representation Learning workshops in higher tier conferences (KDD, AAAI, NIPS)?
2. There are fairly good papers in conferences such as ASONAM or WSDM.

Dear author,

We have three papers on graph neural networks that are missed by this repository. They are:
"A Graph-to-Sequence Model for AMR-to-Text Generation" in ACL 18
Bib: https://aclanthology.info/papers/P18-1150/p18-1150.bib

"Sentence-State LSTM for Text Representation" in ACL 18
Bib: https://aclanthology.info/papers/P18-1030/p18-1030.bib

"N-ary Relation Extraction using Graph State LSTM" in EMNLP 18
@inproceedings{sqltotext_emnlp18,
author = {Linfeng Song and
Yue Zhang and
Zhiguo Wang and
Daniel Gildea},
title = {N-ary Relation Extraction using Graph State {LSTM}},
booktitle = {Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP)},
year = {2018}
}

https://openreview.net/forum?id=39XK7VJ0sKG

Thanks for the awsome project. The followings are some related papers. Thanks

1. Survey
   Title: Hyperbolic Graph Neural Networks: A Review of Methods and Applications
   Arxiv: https://arxiv.org/abs/2202.13852
   Github pages https://github.com/marlin-codes/HGNNs

2. Hyperbolic GNN for temporal network embedding,
   KDD 2021,
   Title: Discrete-time Temporal Network Embedding via Implicit Hierarchical Learning in Hyperbolic Space
   Arxiv: https://arxiv.org/abs/2107.03767
   ACM: https://dl.acm.org/doi/abs/10.1145/3447548.3467422
   Code: https://github.com/marlin-codes/HTGN

3. Overcome oversmoothing problem in dynamic graph,
   ICDM 2020
   Title: FeatureNorm: L2 Feature Normalization for Dynamic Graph Embedding
   https://arxiv.org/abs/2103.00164
   Code: https://github.com/marlin-codes/FeatureNorm-ICDM2020

Great post and collection of papers!

Just want to add a NeurIPS 20 paper to the repo Beta Embeddings for Multi-Hop Logical Reasoning in Knowledge Graphs. We design new set representations with neural logical operations on large heterogeneous KGs. It's the first work that implements all neural first-order logical operations (including negation) over sets. Please feel free to add it to Learning with Sets or Knowledge Graphs.

Besides, our NeurIPS 20 work Graph Information Bottleneck learns an optimal node/graph representation by optimizing variational bounds on mutual information. It demonstrates strong robustness against adversarial attacks. Please feel free to categorize it to adversarial attack/robustness.

Both work are open sourced at BetaE and GIB.

Thanks for your efforts, and really nice work!

Hi!

Thank you for this awesome repository!

Could you please add the following paper, code, and data link to the repository:
Paper: Predicting Dynamic Embedding Trajectory in Temporal Interaction Networks
Authors: Srijan Kumar, Xikun Zhang, Jure Leskovec
Venue: ACM SIGKDD 2019 (Proceedings of the 25th ACM SIGKDD international conference on Knowledge discovery and data mining)
Project page: http://snap.stanford.edu/jodie/
Code: https://github.com/srijankr/jodie/
All datasets: http://snap.stanford.edu/jodie/
Spotlight: https://www.youtube.com/watch?v=ItBmU8681j0

Many thanks,
Srijan

Is a Single Embedding Enough? Learning Node Representations that Capture Multiple Social Contexts

http://epasto.org/papers/www2019splitter.pdf

https://github.com/benedekrozemberczki/Splitter

Hi, thanks for your kindly sharing. I am curious about the "more". For example, when I click the "more" button under NIPS conference, it will show more papers. What's the difference between these papers?

Hi,

Could you please add our ACL 2019 Workshop paper into the repo? In our paper, we propose an end-to-end coreference resolver by combining pre-trained BERT with Relational Graph Convolutional Network (R-GCN).

You could find the paper here: https://arxiv.org/abs/1905.08868, and codes here: https://github.com/ianycxu/RGCN-with-BERT

Best regards,
Yinchuan

WWW2019 paper Heterogeneous Graph Attention Networkhttp://pengcui.thumedialab.com/papers/HeterogeneousGAN.pdf

There are couples of high-quality GNN-based recommendation and information retrieval papers published in SIGIR.