KEYWORDS: Data modeling, Performance modeling, Matrices, Feature extraction, Design and modelling, Proteins, Neural networks, Data processing, Data mining
Autoencoders, as a type of generative self-supervised learning, have received increasing attention in information processing and data mining in recent years. However, existing autoencoders usually generate graph data conforming to the feature distribution from only one aspect of reconstructing edge or node features, which allows the models to extract only a single level of information, limiting their application in real-world applications. In this paper, we propose DummyMAE, a generative self-supervised learning framework that synchronously generates edge and node features. In general, it losslessly converts vertex graphs into corresponding line graphs by introducing edge-to-vertex transformations. The vertex graph provides the model with information on node features, and the line graph provides the model with the ability to capture information on the graph structure, which complements each other. The task of simultaneously reconstructing edges and features is achieved in this way. The task of graph classification serves as a pivotal component within the realm of graph learning, we have conducted sufficient experiments on four widely used graph classification datasets, and the results show that DummyMAE outperforms the current state-of-the-art baselines for the graph classification task.
KEYWORDS: Tunable filters, Social networks, Data modeling, Performance modeling, Matrices, Linear filtering, Digital filtering, Detection theory, Neural networks, Data mining
Graph anomaly detection in graph data has received significant attention due to its practical significance in various vital applications such as network security, finance, and social networks. The current mainstream approach for attribute graph anomaly detection is based on contrastive learning using graph neural networks, which only consider homogeneous low-frequency signals. However, in attribute networks, normal and anomalous nodes exhibit different frequency patterns. This motivates the proposal of a graph anomaly detection framework based on multi-frequency reconstruction to capture the signal patterns of anomaly. Specifically, our method constructs multiple filters based on target nodes and utilizes two modules, namely, low-frequency reconstruction and contrastive learning, for anomaly detection. The generative low-frequency reconstruction module enables us to capture anomalies in the high-frequency attribute space, while the contrastive learning module leverages richer structural information from multiple subgraphs to capture anomalies in the structural and mixed spaces. We conducted extensive experiments on five publicly available datasets, demonstrating that our method significantly outperforms state-of-the-art approaches.
Food adulteration driven by economic interests is an important cause of food safety. Camel milk is widely sought for its high nutritional and medicinal value; some businesses adulterate it for profiteering due to its low yield and high price. Traditional adulteration detection methods rely on supervised learning, which is limited by the data of unknown categories in practical application scenarios, and it is difficult to solve the adulteration problem of category imbalance. For the above scenario, this paper proposes a camel milk adulteration detection framework FIAD based on an unsupervised anomaly detection algorithm, which starts from the perspective of anomaly detection and automatically captures and isolates anomalous features in the data through a tree algorithm without manual labeling of data, directly targeting the adulteration identification problem of category imbalance. We tested the discrimination performance of FIAD in a batch of category-imbalanced camel milk adulteration datasets. At 10% and 20% category imbalance, FIAD achieved AUCs of 0.943 and 0.959 and Recall of 0.915 and 0.949, while occupying less memory, better than eight baseline models. The results show that FIAD has excellent comprehensive identification performance and provides a low-cost and high-efficiency identification method for camel milk adulteration identification.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.