Research of the Evaluation of Customs Generative Artificial Intelligence with Information Entropy
According to a report bythe WCO,the Customs community will increasingly adopt generative artificial intelligence(GAI)technology.Information entropy,as a quantitative evaluation metric,can be used to assess the adaptability of various GAI scenarios and compare alternative technologicalframeworks under different Customs collaboration models.The results demonstrate that employing external economic and trade corpus or general corpus alone,while taking into consideration resource investment and data security,results in limited applicable scenarios,roughly meeting the needs of only one scenario labeled as"investigation and research".However,with the collaboration of internal and external artificial intelligence(AI)resources,particularly by querying external GAI models to obatain domain tags and parameters,not only is the resource investment relatively low,but it can also roughly cover the many scenariosenumerated by the WCO report.With proper fine-tuning,GAI can be integrated into Customs'IT environment,allowing for the creation of new types of productive forces based on GAI.
NETL
NSTL
万方数据
