Graph neural networks (GNNs) have achieved impressive results in various graph learning tasks. Backdoor attacks pose a significant threat to GNNs, with a focus on dirty-label attacks. However, these attacks often necessitate the inclusion of blatantly incorrect inputs into the training set, rendering them easily detectable through simple filtering. In response to this challenge, we introduce Clean-Label Graph Backdoor Attack (CGBA). The majority of features in the generated poisoned nodes align with their true labels, significantly enhancing the difficulty of detecting the attack. Firstly, leveraging the uncertainty inherent in the GNNs, we develop a low-budget strategy for selecting poisoned nodes. This approach focuses on nodes in the target class with uncertain and low-degree classifications, allowing for efficient attacks within a limited budget while mitigating the impact on other clean nodes. Secondly, we present an innovative strategy for generating feature triggers. By boosting the confidence of poisoned samples in the target class, this tactic establishes a robust association between the trigger and the target class, even without modifying the labels of poisoned nodes. Additionally, we incorporate two constraints to reduce disruption to the graph structure. In conclusion, comprehensive experimental results unequivocally showcase CGBA's exceptional attack performance across three benchmark datasets and four GNNs models. Notably, the attack targeting the GraphSAGE model attains a 100% success rate, accompanied by a marginal benign accuracy drop of no more than 0.5%.

To ensure the safety and efficacy of drugs, quality control and supervision of drugs are crucial. However biological drug production is complex, it usually costs much effect to collect and analysis the quality data. This article aims to develop an integrated supervision solution for biopharmaceutical quality, in order to control the quality issues such as deviation of culture medium, and deviation in active ingredients which are easy to occur in the production of biopharmaceuticals. A novel design based on the spectroscopy and machine learning technologies is introduced to collects all quality raw data online. The quality supervision system controls the sampling, inspection, and generation of inspection reports, over traditional IT systems during the production processes. The integrate quality data in system will be used for quality tracing and retrospective analysis, with powerful online data models. Through practical experience in a pharmaceutical company, solution based on the spectroscopy and machine learning improves the efficiency of raw material testing such as culture media, protein solution. The designed architecture of system reduces quality risks during raw material inspections, warehousing, and other biopharmaceutical production processes. The quality data collected in storage, transportation, sales could be monitored and analyzed quickly with online data models. It also enhances the ability to review tendency of all the quality data monthly or annually after biopharmaceutical production, in order to ensure safety of biological drugs.