![](\"https://cdn.analyticsvidhya.com/wp-content/uploads/2018/03/Graph-Theory-768x481.jpg\")
\n \n A graph is a data structure and a universal language for describing relational data structures. A graph consists of a set of entities called nodes and interactions between pairs of those entities known as edges. Graphs represent relational data more efficiently than tabular data or other types of data structures, so graph learning is gaining more attention as a substantial machine learning method.\nDifferent kinds of computational learning and data processing have grown rapidly over the past decades. Machine learning, deep learning, reinforcement learning, etc. have revolutionized our lives, and data science is involved in almost all sciences. As we increase our knowledge about the world, we face more complexity in various fields. We see deep complexity in systems that are composed of numerous entities and their interactions so that they are not easy to understand anymore. We can call them complex networks and explore the main methods to treat different kinds of them.
\n\nIn this work, graph mining, graph learning, and AI in complex networks are briefly reviewed. To be as concise as possible we skip detailed concepts and theories behind graph theory and primarily focus on recent trends that have been so interesting not just in AI but also in various theoretical and practical fields. So, we explore several extraordinary applications and novel innovations in graphs and complex networks.
","metadata":{}},{"cell_type":"markdown","source":"\n| X_e∈R^{mc} |
\nDespite their benefits, graphs are not easy to handle. They can be irregular and different in every characteristic, so some crucial operations like convolutions are tricky to apply to the graph domain. But motivated by CNNs, RNNs, and other successful innovations from deep learning, conceptions of vital operations have been rapidly generalized over the past few years to handle the complexity of graph data.\n\n
\n![](\"https://raw.githubusercontent.com/fedebotu/ICLR2022-OpenReviewData/master/asset/keywords.png\")
\n\n![](\"https://pbs.twimg.com/media/EeSMDD3XgAM0Lsg?format=png&name=900x900\")
\n![](\"https://pbs.twimg.com/media/EFtzXSuWoAAJKGW?format=jpg&name=900x900\")
\nSimulation has always played a significant role in various sciences. However, computational complexity, limited ability to generalize to another area, and difficulty in defining and interpreting physical principles are the main limitations of existing simulations. Recently, graph-based methods have been used in simulation and particle physics domains and have shown successful results.
\n\nThe general method of these works is to define nodes to represent physical particles and edges as interactions between nearby particles. Based on this, a state is defined as a network consisting of nodes and edges at any instant of time, and by applying a set of physical laws and graph-based interactions, each state is transitioned into a state to the next time step. By continuing this process, the initial state ends up in the final state through a series of states. The graph-based simulation method provides simplicity and the ability to generalize some physical properties. Also, network interpretability can be valuable to understand unknown principles. Encoder and decoder are used to transfer between the simulation space and graphic space, as well as a processor model that changes the dynamics of the graph.
\n\n![](\"https://user-images.githubusercontent.com/544269/88065293-60aaba80-cba7-11ea-8f1e-ce3b30774775.png\")
[1] Sanchez-Gonzalez, Alvaro, Jonathan Godwin, Tobias Pfaff, Rex Ying, Jure Leskovec, and Peter Battaglia. \"Learning to simulate complex physics with graph networks.\" In International conference on machine learning, pp. 8459-8468. PMLR, 2020.
\n[2] Pfaff, Tobias, Meire Fortunato, Alvaro Sanchez-Gonzalez, and Peter W. Battaglia. \"Learning mesh-based simulation with graph networks.\" arXiv preprint arXiv:2010.03409 (2020).
\n[3] Shlomi, Jonathan, Peter Battaglia, and Jean-Roch Vlimant. \"Graph neural networks in particle physics.\" Machine Learning: Science and Technology 2, no. 2 (2020): 021001.
\n[4] Belbute-Peres, Filipe De Avila, Thomas Economon, and Zico Kolter. \"Combining differentiable PDE solvers and graph neural networks for fluid flow prediction.\" In international conference on machine learning, pp. 2402-2411. PMLR, 2020.
\nIn recent years, graph-based anomaly detection technique has emerged as a beneficial approach for analyzing relations and connectivity patterns in networks to identify unusual patterns. These methods are becoming undeniable and crucial components in insurance, telecommunication, banking, online credit applications, anti-money laundering, retail holding, trading, and internal organizational fraud detection.
\n \nAs data labels are usually in short supply in many real-world problems, the definition of anomaly as a very first step is a challenging requirement in this field. Also, an anomaly can be relevant to nodes, edges, subgraphs, and events of dynamic graphs, and in every case, describing anomaly features is a very critical task. It's especially essential to consider that organizations and governments are facing an enormous volume of data. For example, hundreds of millions of transactions or phone calls are captured every day, and treating them as network data and applying network-based methods to them, makes it even harder to handle them. Thus, graph mining concepts should meet many data engineering and big data concepts in real-world applications.
\n \nAnother key challenge is security and privacy, as these massive datasets contain very hazardous and private information about individuals, companies, and organizations. Therefore, it's necessary to consider state-of-the-art methods of federated learning and decentralized and safe techniques in future works.
\n \nAlbeit economic and financial networks are the main fields of applying these techniques for fraud detection, many more kinds of networks in infrastructure or IOT have been studied by these anomaly detection methods.
\n| | |||||
| | |||||
| |
[1] Ma, Xiaoxiao, Jia Wu, Shan Xue, Jian Yang, Chuan Zhou, Quan Z. Sheng, Hui Xiong, and Leman Akoglu. \"A comprehensive survey on graph anomaly detection with deep learning.\" IEEE Transactions on Knowledge and Data Engineering (2021).
\n[2] Pourhabibi, Tahereh, Kok-Leong Ong, Booi H. Kam, and Yee Ling Boo. \"Fraud detection: A systematic literature review of graph-based anomaly detection approaches.\" Decision Support Systems 133 (2020): 113303.
\n[3] Weber, Mark, Giacomo Domeniconi, Jie Chen, Daniel Karl I. Weidele, Claudio Bellei, Tom Robinson, and Charles E. Leiserson. \"Anti-money laundering in bitcoin: Experimenting with graph convolutional networks for financial forensics.\" arXiv preprint arXiv:1908.02591 (2019).
\n[4] Liu, Yixin, Zhao Li, Shirui Pan, Chen Gong, Chuan Zhou, and George Karypis. \"Anomaly detection on attributed networks via contrastive self-supervised learning.\" IEEE transactions on neural networks and learning systems 33, no. 6 (2021): 2378-2392.
\n\nAs artificial intelligence is a key tool for solving increasing challenges in transport and traffic, network-based methods are gaining popularity in recent years. The main approach is forming graphs of roads and hubs and then using various concepts of network analysis like centrality, flow, connectivity and cut. These applications are not limited to ground transportation and have recently been used to analyze airlines and airports or shipping routes between ports by means of modeling with complex networks.
\n\n[1] Zheng, Chuanpan, Xiaoliang Fan, Cheng Wang, and Jianzhong Qi. \"Gman: A graph multi-attention network for traffic prediction.\" In Proceedings of the AAAI conference on artificial intelligence, vol. 34, no. 01, pp. 1234-1241. 2020.
\n[2] Sun, Xiaoqian, Sebastian Wandelt, and Anming Zhang. \"How did COVID-19 impact air transportation? A first peek through the lens of complex networks.\" Journal of Air Transport Management 89 (2020): 101928.
\n[3] Chen, Weiqi, Ling Chen, Yu Xie, Wei Cao, Yusong Gao, and Xiaojie Feng. \"Multi-range attentive bicomponent graph convolutional network for traffic forecasting.\" In Proceedings of the AAAI conference on artificial intelligence, vol. 34, no. 04, pp. 3529-3536. 2020.\n
\n[4] Fang, Zheng, Qingqing Long, Guojie Song, and Kunqing Xie. \"Spatial-temporal graph ode networks for traffic flow forecasting.\" In Proceedings of the 27th ACM SIGKDD conference on knowledge discovery & data mining, pp. 364-373. 2021.
\n[5] Zhang, Qin, Keping Yu, Zhiwei Guo, Sahil Garg, Joel JPC Rodrigues, Mohammad Mehedi Hassan, and Mohsen Guizani. \"Graph neural network-driven traffic forecasting for the connected internet of vehicles.\" IEEE Transactions on Network Science and Engineering 9, no. 5 (2021): 3015-3027.
\n[6] Sui, Zhongyi, Yamin Huang, Yuanqiao Wen, Chunhui Zhou, and Xi Huang. \"Marine traffic profile for enhancing situational awareness based on complex network theory.\" Ocean Engineering 241 (2021): 110049.\n\n
\n\nDeepMind describes the use of graphs in protein folding problem as follows\n
\n \nA folded protein can be thought of as a “spatial graph”, where residues are the nodes and edges connect the residues in close proximity. This graph is important for understanding the physical interactions within proteins, as well as their evolutionary history. For the latest version of AlphaFold, used at CASP14, we created an attention-based neural network system, trained end-to-end, that attempts to interpret the structure of this graph, while reasoning over the implicit graph that it’s building. It uses evolutionarily related sequences, multiple sequence alignment (MSA), and a representation of amino acid residue pairs to refine this graph.\n\n\n DeepMind\n :Team, AlphaFold. \"AlphaFold: A Solution to a 50-year-old grand challenge in biology.\" DeepMind, November 30 (2020).","metadata":{}},{"cell_type":"markdown","source":"\n# Cheminformatics and Computational Materials Science\n\n\n
![](\"https://i.ibb.co/68hnp72/cheminformatics1.png\")
\n![](\"https://media.springernature.com/w440/springer-static/cover-hires/journal/43588/1/1\")
One of the most crucial parts of almost every task in cheminformatics and computational drug or material design is the representation of molecules or chemical structures. Although there are several common text-based methods like ‘SMILES’, their resulting representations called “fingerprints” can’t encode all structural information of a 3D structure. This task can be a natural fit for graphs as they can offer valuable insights into the structure of nodes that correspond to atoms and edges that correspond to bonds. So, applying an extended variety of graph neural networks and methods GNNs can learn how to represent atoms based on their local neighbor and available properties of atoms and bonds. Also, this general field involves many tasks with different characteristics. For example, small molecule drugs, polymers, and crystals each require specific attention in future works related to graphs.\n
\n[1] Gilmer, Justin, Samuel S. Schoenholz, Patrick F. Riley, Oriol Vinyals, and George E. Dahl. \"Neural message passing for quantum chemistry.\" In International conference on machine learning, pp. 1263-1272. PMLR, 2017.
\n[2] Schmidt, Jonathan, Mário RG Marques, Silvana Botti, and Miguel AL Marques. \"Recent advances and applications of machine learning in solid-state materials science.\" npj Computational Materials 5, no. 1 (2019): 83.
\n[3] Kearnes, Steven, Kevin McCloskey, Marc Berndl, Vijay Pande, and Patrick Riley. \"Molecular graph convolutions: moving beyond fingerprints.\" Journal of computer-aided molecular design 30 (2016): 595-608.
\n[4] You, Jiaxuan, Bowen Liu, Zhitao Ying, Vijay Pande, and Jure Leskovec. \"Graph convolutional policy network for goal-directed molecular graph generation.\" Advances in neural information processing systems 31 (2018).
\n[5] Chen, Chi, Weike Ye, Yunxing Zuo, Chen Zheng, and Shyue Ping Ong. \"Graph networks as a universal machine learning framework for molecules and crystals.\" Chemistry of Materials 31, no. 9 (2019): 3564-3572.
\n[6] Rong, Yu, Yatao Bian, Tingyang Xu, Weiyang Xie, Ying Wei, Wenbing Huang, and Junzhou Huang. \"Self-supervised graph transformer on large-scale molecular data.\" Advances in Neural Information Processing Systems 33 (2020): 12559-12571.
\n[7] Jiang, Dejun, Zhenxing Wu, Chang-Yu Hsieh, Guangyong Chen, Ben Liao, Zhe Wang, Chao Shen, Dongsheng Cao, Jian Wu, and Tingjun Hou. \"Could graph neural networks learn better molecular representation for drug discovery? A comparison study of descriptor-based and graph-based models.\" Journal of cheminformatics 13, no. 1 (2021): 1-23.
\n[8] Wieder, Oliver, Stefan Kohlbacher, Mélaine Kuenemann, Arthur Garon, Pierre Ducrot, Thomas Seidel, and Thierry Langer. \"A compact review of molecular property prediction with graph neural networks.\" Drug Discovery Today: Technologies 37 (2020): 1-12.
\n[9] Fang, Xiaomin, Lihang Liu, Jieqiong Lei, Donglong He, Shanzhuo Zhang, Jingbo Zhou, Fan Wang, Hua Wu, and Haifeng Wang. \"Geometry-enhanced molecular representation learning for property prediction.\" Nature Machine Intelligence 4, no. 2 (2022): 127-134.
\n\n\n\n\n\n\n","metadata":{}},{"cell_type":"markdown","source":"\n# Brain and Computational Neuroscience\n\n![](\"https://pbs.twimg.com/media/Bk2ujsoCIAAQuAq?format=jpg&name=medium\")
Contrary to the gradual emergence of complex networks, mostly artificial and conceptual, one of the most complex unknown networks has always existed at a distance much closer to humans. The brain is a complex organ consisting of billions of neurons, whose outstanding capabilities and behaviors are still being studied by scientists. Modeling the brain as a complex network of connections between sets of neurons that have their tasks and behaviors can be a new tool in neuroscience and its diverse and growing branches.
\n\n[2] lists some neurological disorders that have been studiesused by graph theory:\n\n\nGraph theory methods, when applied properly, can offer important new insights into the structure and function of networked brain systems, including their architecture, evolution, development, and clinical disorders.\n
\n
...
\nAmong the most widely encountered and biologically meaningful aspects of brain networks is their organization into distinct network communities or modules. Modules are useful to partition larger networks into basic “building blocks,” ie, internally densely connected clusters that are more weakly interconnected amongst each other.\n
...
\nOne of the most widely used approaches in network neuroscience is modularity maximization which aims to divide a given network into a set of nonoverlapping communities by maximizing a global objective function.\n
...
\nModularity maximization faces important methodological limitations that need to be addressed. One limitation refers to the existence of several, often quite numerous, distinct partitions that are almost equally optimal.\n
...
\nAnother fundamental limitation is the inability of the original modularity metric to detect modules below a certain size.\n
...
\nWhile modularity maximization detects nonoverlapping communities, it may also be useful to define \nmodules as overlapping communities.
Sporns, Olaf. \"Graph theory methods: applications in brain networks.\" Dialogues in clinical neuroscience (2022).
\n\n\n\nTo create these graphs, nodes are defined as brain regions of interest (ROIs) and edges are defined as the functional connectivity between those ROIs, computed as the pairwise correlations of functional magnetic resonance imaging (fMRI) time series.
\n
Li, Xiaoxiao, Yuan Zhou, Nicha Dvornek, Muhan Zhang, Siyuan Gao, Juntang Zhuang, Dustin Scheinost, Lawrence H. Staib, Pamela Ventola, and James S. Duncan. \"Braingnn: Interpretable brain graph neural network for fmri analysis.\" Medical Image Analysis 74 (2021): 102233.
\n\n[1] Sporns, Olaf. \"Graph theory methods: applications in brain networks.\" Dialogues in clinical neuroscience (2022).
\n[2] Farahani, Farzad V., Waldemar Karwowski, and Nichole R. Lighthall. \"Application of graph theory for identifying connectivity patterns in human brain networks: a systematic review.\" frontiers in Neuroscience 13 (2019): 585.
\n[3] Li, Xiaoxiao, Yuan Zhou, Nicha Dvornek, Muhan Zhang, Siyuan Gao, Juntang Zhuang, Dustin Scheinost, Lawrence H. Staib, Pamela Ventola, and James S. Duncan. \"Braingnn: Interpretable brain graph neural network for fmri analysis.\" Medical Image Analysis 74 (2021): 102233.
\n\nMany patients take different drugs at the same time and the number of different drug combinations is very high. Also, clinical trials to investigate pharmaceutical challenges are very expensive. Therefore, exploring drug interactions has always been a critical issue in this field due to the risk of adverse drug interactions. In addition, the optimal combination of existing drugs to obtain the desired results significantly reduces the financial and time costs required to prepare new drugs for new or specific diseases. Studying this beneficial aspect of drug interactions gained unprecedented growth during the coronavirus epidemic.
\n\nFor applying graph-based methods, drugs can be represented as nodes and the goal would be predicting edges as interactions. Existing link prediction approaches can be categorized into three families. Heuristic feature-based, latent embedding-based, and neural network-based. Heuristic methods compute node similarity score as the probability of every pair of nodes being connected. There are three levels of these methods. First-order methods like preferential attachment, common neighbors, and Jaccard measure. Second-order methods like Adamic-Adar, resource allocation, and high-order heuristics like SimRank. But these heuristics are usually unable to represent deep structural features. Embedding-based approaches try to learn the latent features of nodes. The most classical method is matrix factorization which takes DDI prediction as a matrix completion task and aims at reconstructing the adjacency matrix. NN-based Methods and a variety of novel deep-learning techniques have been recently applied to explore deep and more abstract structural features.
\n\nIn graph-based drug interaction modeling, the datasets published in different sources in a single network are used, and the chemical characteristics of drugs and interaction characteristics are assigned to the topological features of the graph vertices. Furthermore, other diverse biological networks like protein-protein or gene-protein and gene-disease can be aggregated into the graph models.\nAlthough text inputs are typically represented as a sequence of tokens, a rich variety of NLP problems can be best expressed with a graph structure. As a result, there is a surge of interest in developing new deep-learning techniques on graphs for many NLP tasks.
\n\n[1] Zhou, Yadi, Yuan Hou, Jiayu Shen, Yin Huang, William Martin, and Feixiong Cheng. \"Network-based drug repurposing for novel coronavirus 2019-nCoV/SARS-CoV-2.\" Cell discovery 6, no. 1 (2020): 14.
\n[2] Zitnik, Marinka, Monica Agrawal, and Jure Leskovec. \"Modeling polypharmacy side effects with graph convolutional networks.\" Bioinformatics 34, no. 13 (2018): i457-i466.
\n[3] Morselli Gysi, Deisy, Ítalo Do Valle, Marinka Zitnik, Asher Ameli, Xiao Gan, Onur Varol, Susan Dina Ghiassian et al. \"Network medicine framework for identifying drug-repurposing opportunities for COVID-19.\" Proceedings of the National Academy of Sciences 118, no. 19 (2021): e2025581118.
\n \nAlthough text data is usually represented as a sequence\nof tokens, utilizing graph structures as representation form can be benefitial in many NLP problems. Therefore, there is a surge of interest in innovating graph-base solutions for a large number of NLP problems.
\n\nText classification: \nText classification is a classical task in natural language processing. Constructing a heterogeneous text graph, containing word nodes and document nodes. connectivity between every pair of word nodes is determined using co-occurrence of those words also connectivity between a word node and a document node is determined using word’s document frequency. then text classification problem turns into a heterogeneous graph node classification problem. and strong classification performances reported by applying graph convolution even with a small proportion of labeled documents. also This method is suitable for interpretablity of text classification.\n
\n\n\nAs there is a pressing need for tagging and processing news and tweets, short text classification has found rich and critical applications nowadays. But in many real-world applications, labeled training data is in short supply. Hence, this task requires semi-supervised techniques and it would be valuable to take full advantage of enormous unlabeled data in addition to limited labeled data. Since GNNs enable the integration of any additional information in addition to text data, and then the resultant augmented information can be propagated along the graph, they can be beneficial for this task. \n[2]\nconstruct a graph containing three type of nodes including short texts or documents(D), topics(T) and entities(E).\nA document node is linked to all topic nodes it relates to. Entity linking tools recognize entities of documents and entity embeddings are learned by word2vec2 based on the Wikipedia corpus and next a document node is linked to all entity nodes it contains. Moreover entity nodes connect to each other if the cosine similarity score of their embeddings is above a predefined threshold.\nIn this framework, the integration of topics and entities enriches the semantics of the short texts and improves the performance of short-text classification by applying graph-based methods on resultant graph.\n
\n\n\nText Summarization: [4] \nThe proposed method constructs a graph with sentences as the nodes and similarity between two sentences as the weight of the edge between them.\nThis method can be classified among extractive summarization methods that try to select the best combination of sentences that gives the principal information of the original text. Then target summary can be formed by combining the top n most substantial sentences. The optimal and best collection of sentences should be selected to maximize overall importance while minimizing redundancy.\nIn this method, input text decomposes into a set of sentences that it contains, and then, each sentence is tokenized into a collection of words and similarity among the sentences is defined by those word vectors.\nNext, a complete weighted graph G = (V, E, W) is formed, where V is the set of nodes corresponding to sentences represented by the word vector, E is the set of edges and W is the set of weights assigned to the edges of the graph G. The weight w associated with an edge is determined using text-based techniques.\nThe consequent graph represents the similarity graph of the input document, but it's obviously a complete graph, thus edges with weights less than a predefined threshold, can be removed.\nFor weight assignment, a score is determine to every node corresponding a sentence. \nFinally, for constructing the summary, the top n sentences with the highest scores are selected and arranged according to the order that they appear in the input text. \n
\n\n\n[3] provides a comprehensive review of Graph Neural Networks for Natural\nLanguage Processing and mentions these applications as the most important tasks of NLP that can be improved by graph-baed approaches: Natural Language Generation, Machine Reading Comprehension and Question Answering, Dialog Systems, Text Classification, Text Matching, Topic Modeling, Sentiment Classification, Knowledge Graph, Information Extraction, Semantic and Syntactic Parsing, Reasoning, Semantic Role Labelling.
\n\n[1] Yao, Liang, Chengsheng Mao, and Yuan Luo. \"Graph convolutional networks for text classification.\" In Proceedings of the AAAI conference on artificial intelligence, vol. 33, no. 01, pp. 7370-7377. 2019.
\n[2] Huang, Lianzhe, Dehong Ma, Sujian Li, Xiaodong Zhang, and Houfeng Wang. \"Text level graph neural network for text classification.\" arXiv preprint arXiv:1910.02356 (2019).
\n[3] Linmei, Hu, Tianchi Yang, Chuan Shi, Houye Ji, and Xiaoli Li. \"Heterogeneous graph attention networks for semi-supervised short text classification.\" In Proceedings of the 2019 conference on empirical methods in natural language processing and the 9th international joint conference on natural language processing (EMNLP-IJCNLP), pp. 4821-4830. 2019.
\n[4] Koncel-Kedziorski, Rik, Dhanush Bekal, Yi Luan, Mirella Lapata, and Hannaneh Hajishirzi. \"Text generation from knowledge graphs with graph transformers.\" arXiv preprint arXiv:1904.02342 (2019).
\n[5] Wu, Lingfei, Yu Chen, Kai Shen, Xiaojie Guo, Hanning Gao, Shucheng Li, Jian Pei, and Bo Long. \"Graph neural networks for natural language processing: A survey.\" Foundations and Trends® in Machine Learning 16, no. 2 (2023): 119-328.
\n[6] Lu, Zhibin, Pan Du, and Jian-Yun Nie. \"VGCN-BERT: augmenting BERT with graph embedding for text classification.\" In Advances in Information Retrieval: 42nd European Conference on IR Research, ECIR 2020, Lisbon, Portugal, April 14–17, 2020, Proceedings, Part I 42, pp. 369-382. Springer International Publishing, 2020.
\n[7] Mallick, Chirantana, Ajit Kumar Das, Madhurima Dutta, Asit Kumar Das, and Apurba Sarkar. \"Graph-based text summarization using modified TextRank.\" In Soft Computing in Data Analytics: Proceedings of International Conference on SCDA 2018, pp. 137-146. Springer Singapore, 2019.
\n \nAs robotics improves, more robots appear in every environment and robots must interact with each other and humans too. So in multi-robot systems, collision-free and efficient navigation is a fundamental requirement and an active field of research. This problem which is generally referred to as Multi-Robot Path Planning or Multi-Agent Path Finding can be studied by graph-based methods.
\n\nIn many real-world implementations, robots are restricted in observation and communication range. These scenarios impose a decentralized structure, where each robot has only partial information and separately computes its trajectory. It's very noteworthy that in this way, the resultant performance is not optimal in contrast with centralized methods instead these methods are more suitable in practice and also can significantly reduce the computational complexity of the planning task.
\n\nBut what information is crucial to be shared and how and when it must be shared among robots? These questions are not well addressed by current machine learning approaches yet. Constructing a network involving nodes as robots and edges as communication links can be helpful for modeling the collective robot system and its communication network. Then it's possible to utilize Graph Neural Networks and achieve a framework that can learn what information is relevant to be shared between robots and the whole team.
\n\n\n\n
\n\n\n\n
\n\n\n
\n\n[1] Li, Qingbiao, Fernando Gama, Alejandro Ribeiro, and Amanda Prorok. \"Graph neural networks for decentralized multi-robot path planning.\" In 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 11785-11792. IEEE, 2020.
\n[2] Li, Qingbiao, Weizhe Lin, Zhe Liu, and Amanda Prorok. \"Message-aware graph attention networks for large-scale multi-robot path planning.\" IEEE Robotics and Automation Letters 6, no. 3 (2021): 5533-5540.
\n[3] Tolstaya, Ekaterina, Fernando Gama, James Paulos, George Pappas, Vijay Kumar, and Alejandro Ribeiro. \"Learning decentralized controllers for robot swarms with graph neural networks.\" In Conference on robot learning, pp. 671-682. PMLR, 2020.
\n[4] Chen, Yuying, Congcong Liu, Bertram E. Shi, and Ming Liu. \"Robot navigation in crowds by graph convolutional networks with attention learned from human gaze.\" IEEE Robotics and Automation Letters 5, no. 2 (2020): 2754-2761.
\n[5] Dang, Tung, Frank Mascarich, Shehryar Khattak, Christos Papachristos, and Kostas Alexis. \"Graph-based path planning for autonomous robotic exploration in subterranean environments.\" In 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 3105-3112. IEEE, 2019.
\n[6] Liang, Ming, Bin Yang, Rui Hu, Yun Chen, Renjie Liao, Song Feng, and Raquel Urtasun. \"Learning lane graph representations for motion forecasting.\" In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part II 16, pp. 541-556. Springer International Publishing, 2020.
\n[7] Wang, Zheyuan, and Matthew Gombolay. \"Learning scheduling policies for multi-robot coordination with graph attention networks.\" IEEE Robotics and Automation Letters 5, no. 3 (2020): 4509-4516.
\n[8] Shi, Liushuai, Le Wang, Chengjiang Long, Sanping Zhou, Mo Zhou, Zhenxing Niu, and Gang Hua. \"SGCN: Sparse graph convolution network for pedestrian trajectory prediction.\" In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 8994-9003. 2021.
\n \n[1] Wang, Yue, Yongbin Sun, Ziwei Liu, Sanjay E. Sarma, Michael M. Bronstein, and Justin M. Solomon. \"Dynamic graph cnn for learning on point clouds.\" ACM Transactions on Graphics (tog) 38, no. 5 (2019): 1-12.
\n[2] Wu, Shu, Yuyuan Tang, Yanqiao Zhu, Liang Wang, Xing Xie, and Tieniu Tan. \"Session-based recommendation with graph neural networks.\" In Proceedings of the AAAI conference on artificial intelligence, vol. 33, no. 01, pp. 346-353. 2019.
\n[3] Wu, Zonghan, Shirui Pan, Guodong Long, Jing Jiang, Xiaojun Chang, and Chengqi Zhang. \"Connecting the dots: Multivariate time series forecasting with graph neural networks.\" In Proceedings of the 26th ACM SIGKDD international conference on knowledge discovery & data mining, pp. 753-763. 2020.
\n[4] Jones, Payton J., Ruofan Ma, and Richard J. McNally. \"Bridge centrality: A network approach to understanding comorbidity.\" Multivariate behavioral research 56, no. 2 (2021): 353-367.
\n[5] Robinaugh, Donald J., Ria HA Hoekstra, Emma R. Toner, and Denny Borsboom. \"The network approach to psychopathology: a review of the literature 2008–2018 and an agenda for future research.\" Psychological medicine 50, no. 3 (2020): 353-366.
\n[6] Borsboom, Denny, Marie K. Deserno, Mijke Rhemtulla, Sacha Epskamp, Eiko I. Fried, Richard J. McNally, Donald J. Robinaugh et al. \"Network analysis of multivariate data in psychological science.\" Nature Reviews Methods Primers 1, no. 1 (2021): 58.
\n[7] Cappart, Quentin, Didier Chételat, Elias B. Khalil, Andrea Lodi, Christopher Morris, and Petar Velickovic. \"Combinatorial Optimization and Reasoning with Graph Neural Networks.\" In IJCAI, pp. 4348-4355. 2021.
\n[8] Wang, Wenhan, Ge Li, Bo Ma, Xin Xia, and Zhi Jin. \"Detecting code clones with graph neural network and flow-augmented abstract syntax tree.\" In 2020 IEEE 27th International Conference on Software Analysis, Evolution and Reengineering (SANER), pp. 261-271. IEEE, 2020.
\n[9] Kapoor, Amol, Xue Ben, Luyang Liu, Bryan Perozzi, Matt Barnes, Martin Blais, and Shawn O'Banion. \"Examining covid-19 forecasting using spatio-temporal graph neural networks.\" arXiv preprint arXiv:2007.03113 (2020).
\n[10] Liu, Fucheng, Yu Wen, Dongxue Zhang, Xihe Jiang, Xinyu Xing, and Dan Meng. \"Log2vec: A heterogeneous graph embedding based approach for detecting cyber threats within enterprise.\" In Proceedings of the 2019 ACM SIGSAC conference on computer and communications security, pp. 1777-1794. 2019.
\n[11] Li, Wei, Ruihan Bao, Keiko Harimoto, Deli Chen, Jingjing Xu, and Qi Su. \"Modeling the stock relation with graph network for overnight stock movement prediction.\" In Proceedings of the twenty-ninth international conference on international joint conferences on artificial intelligence, pp. 4541-4547. 2021.
\n[12] Almasan, Paul, José Suárez-Varela, Krzysztof Rusek, Pere Barlet-Ros, and Albert Cabellos-Aparicio. \"Deep reinforcement learning meets graph neural networks: Exploring a routing optimization use case.\" Computer Communications 196 (2022): 184-194.
\n\ninnovating state-of-the-art graph neural networks both general and domain-specific in addition to concepts like graph embedding and transfer learning are active fields of research.
\n\n[1] Velickovic, Petar, Guillem Cucurull, Arantxa Casanova, Adriana Romero, Pietro Lio, and Yoshua Bengio. \"Graph attention networks.\" stat 1050, no. 20 (2017): 10-48550.
\n[2] Hamilton, Will, Zhitao Ying, and Jure Leskovec. \"Inductive representation learning on large graphs.\" Advances in neural information processing systems 30 (2017).
\n[3] Xu, Keyulu, Weihua Hu, Jure Leskovec, and Stefanie Jegelka. \"How powerful are graph neural networks?.\" arXiv preprint arXiv:1810.00826 (2018).
\n[4] Schlichtkrull, Michael, Thomas N. Kipf, Peter Bloem, Rianne Van Den Berg, Ivan Titov, and Max Welling. \"Modeling relational data with graph convolutional networks.\" In The Semantic Web: 15th International Conference, ESWC 2018, Heraklion, Crete, Greece, June 3–7, 2018, Proceedings 15, pp. 593-607. Springer International Publishing, 2018.
\n[5] Xu, Keyulu, Chengtao Li, Yonglong Tian, Tomohiro Sonobe, Ken-ichi Kawarabayashi, and Stefanie Jegelka. \"Representation learning on graphs with jumping knowledge networks.\" In International conference on machine learning, pp. 5453-5462. PMLR, 2018.
\n[6] Zhang, Muhan, and Yixin Chen. \"Link prediction based on graph neural networks.\" Advances in neural information processing systems 31 (2018).
\n[7] You, Yuning, Tianlong Chen, Yongduo Sui, Ting Chen, Zhangyang Wang, and Yang Shen. \"Graph contrastive learning with augmentations.\" Advances in neural information processing systems 33 (2020): 5812-5823.
\n[8] Chen, Ming, Zhewei Wei, Zengfeng Huang, Bolin Ding, and Yaliang Li. \"Simple and deep graph convolutional networks.\" In International conference on machine learning, pp. 1725-1735. PMLR, 2020.
\n[9] Morris, Christopher, Martin Ritzert, Matthias Fey, William L. Hamilton, Jan Eric Lenssen, Gaurav Rattan, and Martin Grohe. \"Weisfeiler and leman go neural: Higher-order graph neural networks.\" In Proceedings of the AAAI conference on artificial intelligence, vol. 33, no. 01, pp. 4602-4609. 2019.
\n[10] Hu, Weihua, Bowen Liu, Joseph Gomes, Marinka Zitnik, Percy Liang, Vijay Pande, and Jure Leskovec. \"Strategies for pre-training graph neural networks.\" arXiv preprint arXiv:1905.12265 (2019).
\n[11] Li, Guohao, Matthias Muller, Ali Thabet, and Bernard Ghanem. \"Deepgcns: Can gcns go as deep as cnns?.\" In Proceedings of the IEEE/CVF international conference on computer vision, pp. 9267-9276. 2019.
\n[12] Zeng, Hanqing, Hongkuan Zhou, Ajitesh Srivastava, Rajgopal Kannan, and Viktor Prasanna. \"Graphsaint: Graph sampling based inductive learning method.\" arXiv preprint arXiv:1907.04931 (2019).
\n[13] Zhu, Jiong, Yujun Yan, Lingxiao Zhao, Mark Heimann, Leman Akoglu, and Danai Koutra. \"Beyond homophily in graph neural networks: Current limitations and effective designs.\" Advances in neural information processing systems 33 (2020): 7793-7804.
\n[14] Brody, Shaked, Uri Alon, and Eran Yahav. \"How attentive are graph attention networks?.\" arXiv preprint arXiv:2105.14491 (2021).
\n[15] Alon, Uri, and Eran Yahav. \"On the bottleneck of graph neural networks and its practical implications.\" arXiv preprint arXiv:2006.05205 (2020).
\n[16] Dwivedi, Vijay Prakash, and Xavier Bresson. \"A generalization of transformer networks to graphs.\" arXiv preprint arXiv:2012.09699 (2020).
\n \n\n[1] Wu, Zonghan, Shirui Pan, Fengwen Chen, Guodong Long, Chengqi Zhang, and S. Yu Philip. \"A comprehensive survey on graph neural networks.\" IEEE transactions on neural networks and learning systems 32, no. 1 (2020): 4-24.
\n[2] Ji, Shaoxiong, Shirui Pan, Erik Cambria, Pekka Marttinen, and S. Yu Philip. \"A survey on knowledge graphs: Representation, acquisition, and applications.\" IEEE transactions on neural networks and learning systems 33, no. 2 (2021): 494-514.
\n[3] Zhang, Ziwei, Peng Cui, and Wenwu Zhu. \"Deep learning on graphs: A survey.\" IEEE Transactions on Knowledge and Data Engineering 34, no. 1 (2020): 249-270.
\n[4] Goyal, Palash, and Emilio Ferrara. \"Graph embedding techniques, applications, and performance: A survey.\" Knowledge-Based Systems 151 (2018): 78-94.
\n[5] Wu, Lingfei, Peng Cui, Jian Pei, Liang Zhao, and Xiaojie Guo. \"Graph neural networks: foundation, frontiers and applications.\" In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pp. 4840-4841. 2022.
\n \n \n![](\"https://m.media-amazon.com/images/I/81oz+bjUNoL._AC_UF894,1000_QL80_.jpg\")
\n![](\"https://m.media-amazon.com/images/I/513GGCczuXL.jpg\")
\n![](\"https://m.media-amazon.com/images/I/71XXR+bJRcL._AC_UF894,1000_QL80_.jpg\")
\n[1] Barabási, Albert-László. \"Network science.\" Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 371, no. 1987 (2013): 20120375.
\n[2] Easley, David, and Jon Kleinberg. Networks, crowds, and markets: Reasoning about a highly connected world. Vol. 1. Cambridge: Cambridge university press, 2010.
\n[3] Hamilton, William L. Graph representation learning. Morgan & Claypool Publishers, 2020.
","metadata":{}},{"cell_type":"markdown","source":"\n\nThese two collections offer benchmark datasets from different fields and can be good starting points
\n\n[1] Hu, Weihua, Matthias Fey, Marinka Zitnik, Yuxiao Dong, Hongyu Ren, Bowen Liu, Michele Catasta, and Jure Leskovec. \"Open graph benchmark: Datasets for machine learning on graphs.\" Advances in neural information processing systems 33 (2020): 22118-22133.
\n[2] Morris, Christopher, Nils M. Kriege, Franka Bause, Kristian Kersting, Petra Mutzel, and Marion Neumann. \"Tudataset: A collection of benchmark datasets for learning with graphs.\" arXiv preprint arXiv:2007.08663 (2020).
\n","metadata":{}},{"cell_type":"code","source":"f2 = IFrame('https://ogb.stanford.edu/',width='100%', height=450)\ndisplay(f2)","metadata":{"execution":{"iopub.status.busy":"2023-07-16T23:57:17.500861Z","iopub.execute_input":"2023-07-16T23:57:17.501290Z","iopub.status.idle":"2023-07-16T23:57:17.513911Z","shell.execute_reply.started":"2023-07-16T23:57:17.501219Z","shell.execute_reply":"2023-07-16T23:57:17.512499Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"f3 = IFrame('https://chrsmrrs.github.io/datasets/docs/datasets/',width='100%', height=300)\ndisplay(f3)","metadata":{"execution":{"iopub.status.busy":"2023-07-16T23:57:17.517707Z","iopub.execute_input":"2023-07-16T23:57:17.518076Z","iopub.status.idle":"2023-07-16T23:57:17.529257Z","shell.execute_reply.started":"2023-07-16T23:57:17.518048Z","shell.execute_reply":"2023-07-16T23:57:17.528074Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"![link2](\"https://github.com/fedebotu/ICLR2022-OpenReviewData/blob/master/asset/keywords.png\"){kind=link}