In terms of the improvement of efficiency and data maintenance of the ChemiCloud platform, we added government data transfer platform (T-Road) to transmit electronic invoice data, expanded two data systems and provided 2 APIs to National Fire Agency and Industrial Development Administration respectively. Additionally, in conjunction with the reorganization of the Ministry of Environment and other agencies, the platform information presentation had been adjusted. Finally, we continued to maintain fifty-three chemical substance management information systems. A total of three hundred and twelve ChemiCloud accounts were created. We also continued to carry out platform information security, operation and maintenance. Last but not least, we completed three employee trainings, ten meetings with authorities concerned and one cross-functional meeting. In terms of platform security protection and future overall architecture improvement, we completed complete the Ministry of Interior’s Citizen Digital Certificate login system and the online application mechanism for chemical cloud accounts. In addition, the pre-made form was used to optimize the query speed, the permission sharing form design was expanded, and the role recognition permission mechanism was added to optimize the control principle of permissions. Finally, Three component optimizations and the integration specifications of ChemiCloud system interfaces were completed. In terms of big data analysis, the blockchain part had completed the development of digital license management that included sharing functions, providing manufacturers with certificate sharing links, allowing third parties to view toxic chemical substance license information. We also invited three manufacturers to conduct trials and provide feedback. In the part of the high-risk abnormal manufacturer detection model, the best model combination is Variational Graph AutoEncoder (VGAE) and self-supervised contrastive anomaly detection (SCAD), adding to use the anomaly scores for sorting. Furthermore, adding the financial and tax information had significantly improved the model's ability to identify real abnormal samples. In the news crawler part, with the help of ChatGPT tags, a small data set was created, and a food safety emergency module and a keyword extraction module were developed. Therefore, we optimized the value-added promotion of food safety news through refined food safety emergency classification identification and food safety news keyword identification. In the part of extensive collection of chemical substance control information and understanding of management trends, we completed the compilation of three data sets from the Stockholm Convention, the food additive list announced by the United Nations Food Standards Commission (Codex), and Taiwan Environmental Hormones Management Plan Table. We reviewed and updated relevant domestic and foreign regulations list of chemical substances of concern regularly, and the collection of 2023 domestic food-related inspection and sampling results was completed.

Chemical Substance, Inter-Ministerial Cooperation, Alert and Quick Report