Machine Learning applications for international trade

国际贸易中的机器学习应用

• 批准号: 10035345
• 金额: $ 34.24万
• 依托单位: SIFTY ML LTD
• 依托单位国家: 英国
• 项目类别: Collaborative R&D
• 财政年份: 2022
• 资助国家: 英国
• 起止时间: 2022 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=10035345
• 关键词: Machine; Learning; applications; international; trade

Global challenges and shifting trends in e-commerce mean that managing the flow of international goods is becoming increasingly demanding and complex. The rapid growth of freight to local fulfilment centres for global sales platforms such as Amazon and AliExpress exacerbates this. Across the fragmented global logistics industry, customs processing costs are spiralling, and delays are becoming endemic. Trading companies, customs brokers, and freight forwarders urgently need solutions which remove the bottlenecks from their customs processes that cost the global industry billions of pounds.The market opportunity is enormous: the logistics industry handles \>£13.5tn in goods annually; with UK cargo imports \>£500bn (25% with the EU). International incoming parcels create £3.8bn revenues for courier companies.Core bottlenecks stem from a continuing disconnect between national and international package coding and a lack of digitisation for analysis. Data is difficult to obtain and structure, limiting competitor solution capability.SiftyML's machine learning-powered software provides a revolutionary solution which will disrupt the global logistics industry, with a unique customs data structuring and processing tool which can become the operating system of international trade. The project will transform SiftyML's existing solution into a dynamic system adapting in real-time to worldwide changes in requirements, enabling customs brokers to process parcels accurately and swiftly, addressing critical bottlenecks, and increasing efficiencies and revenues. It will disrupt the state-of-the-art, with advanced machine learning unleashed on a massive dataset of real-time customs data to predict, assess and suggest information that can increase the efficiency of goods moving across borders, radically reducing processing times from weeks to days, days to hours, and hours to seconds. SiftyML's algorithms can enable 95% of straightforward parcels to be processed with minimal human effort, freeing up experienced customs officials to concentrate on the 5% of complex parcels which require their expertise, supported by SiftyML's automation and insights to augment the human process. The market opportunity is enormous: the logistics industry handles \>£13.5 trillion in goods annually; with UK cargo imports \>£500 billion (25% with the EU). International incoming parcels create £3.8 billion in revenues for courier companies, but profit margins are tight with existing bottleneck solutions requiring costly staff time and effort.

电子商务的全球挑战和不断变化的趋势意味着管理国际货物流动的要求越来越高，越来越复杂。亚马逊（Amazon）和全球速卖通（AliExpress）等全球销售平台向当地物流中心运送货物的快速增长加剧了这一点。在分散的全球物流行业，海关处理成本正在螺旋式上升，延误正在成为普遍现象。贸易公司、报关行和货运代理迫切需要解决方案，以消除其海关流程中的瓶颈，这些瓶颈使全球行业损失数十亿英镑。市场机会巨大：物流行业每年处理的货物超过13.5万亿英镑;英国货物进口超过5000亿英镑（欧盟占25%）。国际包裹为快递公司创造了38亿英镑的收入。核心瓶颈来自国内和国际包裹编码之间的持续脱节以及缺乏数字化分析。SiftyML的机器学习软件提供了一个革命性的解决方案，将颠覆全球物流业，其独特的海关数据结构和处理工具可以成为国际贸易的操作系统。该项目将把SiftyML现有的解决方案转变为一个动态系统，实时适应全球需求的变化，使海关经纪人能够准确、快速地处理包裹，解决关键瓶颈，提高效率和收入。它将颠覆最先进的技术，在大量实时海关数据集上释放先进的机器学习，以预测、评估和建议可以提高货物跨境运输效率的信息，从根本上将处理时间从数周缩短到数天，从数天缩短到数小时，从数小时缩短到数秒。SiftyML的算法可以使95%的简单包裹以最少的人力进行处理，从而使经验丰富的海关官员能够专注于5%的复杂包裹，这些包裹需要他们的专业知识，并得到SiftyML的自动化和洞察力的支持，以增强人工流程。市场机会巨大：物流行业每年处理超过13.5万亿英镑的货物;英国货物进口超过5000亿英镑（25%来自欧盟）。国际传入包裹为快递公司创造了38亿英镑的收入，但由于现有的瓶颈解决方案需要花费大量的员工时间和精力，利润率很低。