Graph Classification

Graph pooling is an essential ingredient of Graph Neural Networks (GNNs) in graph classification and regression tasks. For these tasks, different pooling strategies have been proposed to generate a graph-level representation by downsampling and summarizing nodes' features in a graph. However, most existing pooling methods are unable to capture distinguishable structural information effectively. Besides, they are prone to adversarial attacks. In this work, we propose a novel pooling method named as HIBPool where we leverage the Information Bottleneck (IB) principle that optimally balances the expressiveness and robustness of a model to learn representations of input data. Furthermore, we introduce a novel structureaware Discriminative Pooling Readout (DiP-Readout) function to capture the informative local subgraph structures in the graph. Finally, our experimental results show that our model significantly outperforms other state-of-art methods on several graph classification benchmarks and more resilient to featureperturbation attack than existing pooling methods 1 .

In the graph classification problem, given is a family of graphs and a group of different categories, and we aim to classify all the graphs (of the family) into the given categories. Earlier approaches, such as graph kernels and graph embedding techniques have focused on extracting certain features by processing the entire graph. However, real world graphs are complex and noisy and these traditional approaches are computationally intensive. With the introduction of the deep learning framework, there have been numerous attempts to create more efficient classification approaches. We modify a kernel graph convolutional neural network approach, that extracts subgraphs (patches) from the graph using various community detection algorithms. These patches are provided as input to a graph kernel and max pooling is applied. We use different community detection algorithms and a shortest path graph kernel and compare their efficiency and performance. In this paper we compare three methods: a gr...

Text mining and text classification are gaining more and more importance in AI related research fields. Researchers are particularly focused on classification systems, based on structured data (such as sequences or graphs), facing the challenge of synthesizing interpretable models, exploiting gray-box approaches. In this paper, a novel gray-box text classifier is presented. Documents to be classified are split into their constituent words, or tokens. Groups of frequent m tokens (or m-grams) are suitably mined adopting the Granular Computing framework. By fastText algorithm, each token is encoded in a real-valued vector and a custom-based dissimilarity measure, grounded on the Edit family, is designed specifically to deal with m-grams. Through a clustering procedure the most representative m-grams, pertaining the corpus of documents, are extrapolated and arranged into a Symbolic Histogram representation. The latter allows embedding documents in a wellsuited real-valued space in which a standard classifier, such as SVM, can safety operate. Along with the classification procedure, an Evolutionary Algorithm is in charge of performing features selection, which is able to select most relevant symbols-m-grams-for each class. This study shows how symbols can be fruitfully interpreted, allowing an interesting knowledge discovery procedure, in lights with the new requirements of modern explainable AI systems. The effectiveness of the proposed algorithm has been proved through a set of experiments on paper abstracts classification and SMS spam detection.

Granular Computing is a powerful information processing paradigm, particularly useful for the synthesis of pattern recognition systems in structured domains (e.g., graphs or sequences). According to this paradigm, granules of information play the pivotal role of describing the underlying (possibly complex) process, starting from the available data. Under a pattern recognition viewpoint, granules of information can be exploited for the synthesis of semantically sound embedding spaces, where common supervised or unsupervised problems can be solved via standard machine learning algorithms. In this companion paper, we follow our previous paper (Martino et al. in Algorithms 15(5):148, 2022) in the context of comparing different strategies for the automatic synthesis of information granules in the context of graph classification. These strategies mainly differ on the specific topology adopted for subgraphs considered as candidate information granules and the possibility of using or neglec...

Text mining and text classification are gaining more and more importance in AI related research fields. Researchers are particularly focused on classification systems, based on structured data (such as sequences or graphs), facing the challenge of synthesizing interpretable models, exploiting gray-box approaches. In this paper, a novel gray-box text classifier is presented. Documents to be classified are split into their constituent words, or tokens. Groups of frequent m tokens (or m-grams) are suitably mined adopting the Granular Computing framework. By fastText algorithm, each token is encoded in a real-valued vector and a custom-based dissimilarity measure, grounded on the Edit family, is designed specifically to deal with m-grams. Through a clustering procedure the most representative m-grams, pertaining the corpus of documents, are extrapolated and arranged into a Symbolic Histogram representation. The latter allows embedding documents in a wellsuited real-valued space in which a standard classifier, such as SVM, can safety operate. Along with the classification procedure, an Evolutionary Algorithm is in charge of performing features selection, which is able to select most relevant symbols-m-grams-for each class. This study shows how symbols can be fruitfully interpreted, allowing an interesting knowledge discovery procedure, in lights with the new requirements of modern explainable AI systems. The effectiveness of the proposed algorithm has been proved through a set of experiments on paper abstracts classification and SMS spam detection.

Fake News Analysis and Graph Classification on a COVID-19 Twitter Dataset by Kriti Gupta Earlier researches have showed that the spread of fake news through social media can have a huge impact to society and also to individuals in an extremely negative way. In this work we aim to study the spread of fake news compared to real news in a social network. We do that by performing classical social network analysis to discover various characteristics, and formulate the problem as a binary classification, where we have graphs modeling the spread of fake and real news. For our experiments we rely on how news are propagated through a popular social media services such as Twitter during the pandemic caused by the COVID-19 virus. In the past, several other approaches classify news as fake or real by deploying various graph embedding techniques and deep learning techniques. In this project we focus on developing a dataset that contains tweets specific to COVID-19 by performing initially text mining on the content of the tweet. Further, we create graphs of the fake and real news along with their retweets and followers and work on the graphs. We perform social network analysis and compare their characteristics. We study the propagation of fake and real news among users using community detection algorithms on the graphs. Finally, we create a model by deploying the Weisfeiler Lehman graph kernel for graph classification on our labeled dataset. The model is able to predict whether a new article is real or fake based on how the corresponding graph of the retweets and followers are connected.

Fake News Analysis and Graph Classification on a COVID-19 Twitter Dataset by Kriti Gupta Earlier researches have showed that the spread of fake news through social media can have a huge impact to society and also to individuals in an extremely negative way. In this work we aim to study the spread of fake news compared to real news in a social network. We do that by performing classical social network analysis to discover various characteristics, and formulate the problem as a binary classification, where we have graphs modeling the spread of fake and real news. For our experiments we rely on how news are propagated through a popular social media services such as Twitter during the pandemic caused by the COVID-19 virus. In the past, several other approaches classify news as fake or real by deploying various graph embedding techniques and deep learning techniques. In this project we focus on developing a dataset that contains tweets specific to COVID-19 by performing initially text mining on the content of the tweet. Further, we create graphs of the fake and real news along with their retweets and followers and work on the graphs. We perform social network analysis and compare their characteristics. We study the propagation of fake and real news among users using community detection algorithms on the graphs. Finally, we create a model by deploying the Weisfeiler Lehman graph kernel for graph classification on our labeled dataset. The model is able to predict whether a new article is real or fake based on how the corresponding graph of the retweets and followers are connected.

Supérieure (a) graph matching without learning (b) with a learned matching function (c) a learned graph model and its matching Figure 1: Graph learning for matching. Our approach learns a graph model from labeled data to provide the best match to instances of a target class. It shows significant improvement over previous approaches for matching.

Graph pooling is an essential ingredient of Graph Neural Networks (GNNs) in graph classification and regression tasks. For these tasks, different pooling strategies have been proposed to generate a graph-level representation by downsampling and summarizing nodes' features in a graph. However, most existing pooling methods are unable to capture distinguishable structural information effectively. Besides, they are prone to adversarial attacks. In this work, we propose a novel pooling method named as HIBPool where we leverage the Information Bottleneck (IB) principle that optimally balances the expressiveness and robustness of a model to learn representations of input data. Furthermore, we introduce a novel structureaware Discriminative Pooling Readout (DiP-Readout) function to capture the informative local subgraph structures in the graph. Finally, our experimental results show that our model significantly outperforms other state-of-art methods on several graph classification benchmarks and more resilient to featureperturbation attack than existing pooling methods 1 .

We propose a novel unsupervised method that assesses the similarity of two videos on the basis of the estimated relatedness of the objects and their behavior, and provides arguments supporting this assessment. A video is represented as a complete undirected action graph that encapsulates information on the types of objects and the way they (inter)act. The similarity of a pair of videos is estimated based on the bipartite Graph Edit Distance (GED) of the corresponding action graphs. As a consequence, on-top of estimating a quantitative measure of video similarity, our method establishes spatiotemporal correspondences between objects across videos if these objects are semantically related, if/when they interact similarly, or both. We consider this an important step towards explainable assessment of video and action similarity. The proposed method is evaluated on a publicly available dataset on the tasks of activity classification and ranking and is shown to compare favorably to state ...

Many tasks in computer vision and pattern recognition are formulated as graph matching problems. Despite the NP-hard nature of the problem, fast and accurate approximations have led to significant progress in a wide range of applications. Learning graph matching functions from observed data, however, still remains a challenging issue. This paper presents an effective scheme to parametrize a graph model for object matching in a classification context. For this, we propose a representation based on a parametrized model graph, and optimize it to increase a classification rate. Experimental evaluations on real datasets demonstrate the effectiveness (in terms of accuracy and speed) of our approach against graph classification with hand-crafted cost functions.

Granular Computing is a powerful information processing paradigm, particularly useful for the synthesis of pattern recognition systems in structured domains (e.g., graphs or sequences). According to this paradigm, granules of information play the pivotal role of describing the underlying (possibly complex) process, starting from the available data. Under a pattern recognition viewpoint, granules of information can be exploited for the synthesis of semantically sound embedding spaces, where common supervised or unsupervised problems can be solved via standard machine learning algorithms. In this work, we show a comparison between different strategies for the automatic synthesis of information granules in the context of graph classification. These strategies mainly differ on the specific topology adopted for subgraphs considered as candidate information granules and the possibility of using or neglecting the ground-truth class labels in the granulation process. Computational results o...

The classification of graph based objects is an important challenge from a knowledge discovery standpoint and has attracted considerable attention recently. In this paper, we present a probabilistic substructure-based approach for classifying a graphbased dataset. More specifically, we use a frequent subgraph mining algorithm to construct substructure based descriptors and apply the maximum entropy principle to convert the local patterns into a global classification model for graph data. Empirical studies conducted on real world data sets showed that the maximum entropy substructure-based approach often outperforms existing feature vector methods using AdaBoost and Support Vector Machine.

Structural pattern recognition approaches offer the most expressive, convenient, powerful but computational expensive representations of underlying relational information. To benefit from mature, less expensive and efficient state-of-the-art machine learning models of statistical pattern recognition they must be mapped to a low-dimensional vector space. Our method of explicit graph embedding bridges the gap between structural and statistical pattern recognition. We extract the topological, structural and attribute information from a graph and encode numeric details by fuzzy histograms and symbolic details by crisp histograms. The histograms are concatenated to achieve a simple and straightforward embedding of graph into a low-dimensional numeric feature vector. Experimentation on standard public graph datasets shows that our method outperforms the state-of-the-art methods of graph embedding for richly attributed graphs.

Granular Computing is a powerful information processing paradigm for synthesizing advanced pattern recognition systems in non-conventional domains. In this article, a novel procedure for the automatic synthesis of suitable information granules is proposed. The procedure leverages a joint sensitivityvs-specificity score that accounts the meaningfulness of candidate information granules for each class considered in the classification problem at hand. Only statistically relevant granules are retained for a graph embedding procedure towards a geometric space, in which standard classification systems can be used without alterations. Performance tests have been carried out by considering open access datasets of fully labelled graphs with arbitrarily complex nodes and/or edges attributes that, by definition, must rely on inexact graph matching procedures to quantify dissimilarities. Two variants of the procedure are investigated: a standard variant, which aims at automatically finding suitable information granules for solving the classification problem as a whole, and a class-specific metric learning variant, in which the optimization procedure is performed in a class-aware fashion. In the latter case, each class will have its own set of information granules, along with the corresponding parameters defining distinct instances of the dissimilarity measure. Computational results show that the proposed algorithm is able to outperform the vast majority of current approaches for graph classification, while at the same time returning a grey-box model, interpretable by field-experts.

Granular Computing is a powerful information processing paradigm for synthesizing advanced pattern recognition systems in non-conventional domains. In this article, a novel procedure for the automatic synthesis of suitable information granules is proposed. The procedure leverages a joint sensitivity-vs-specificity score that accounts the meaningfulness of candidate information granules for each class considered in the classification problem at hand. Only statistically relevant granules are retained for a graph embedding procedure towards a geometric space, in which standard classification systems can be used without alterations. Performance tests have been carried out by considering open access datasets of fully labelled graphs with arbitrarily complex nodes and/or edges attributes that, by definition, must rely on inexact graph matching procedures to quantify dissimilarities. Two variants of the procedure are investigated: a standard variant, which aims at automatically finding suitable information granules for solving the classification problem as a whole, and a class-specific metric learning variant, in which the optimization procedure is performed in a class-aware fashion. In the latter case, each class will have its own set of information granules, along with the corresponding parameters defining distinct instances of the dissimilarity measure. Computational results show that the proposed algorithm is able to outperform the vast majority of current approaches for graph classification, while at the same time returning a grey-box model, interpretable by field-experts.