课题基金 / 基金详情

Extending Zinia no-code AI platform for Streaming AI and Hyper Explainability for SME lenders to optimize price of credit for sub-prime customers, enabling safe and fair access to credit, meeting the growing demand from households crushed by the pandemic

Extending Zinia no-code AI platform for Streaming AI and Hyper Explainability for SME lenders to optimize price of credit for sub-prime customers, enabling safe and fair access to credit, meeting the growing demand from households crushed by the pandemic
扩展 Zinia 无代码人工智能平台,为中小企业贷方提供流式人工智能和超可解释性,以优化次级贷款客户的信贷价格,实现安全和公平的信贷获取,满足受疫情影响的家庭不断增长的需求
批准号:
10024287
负责人:
金额:
$35.04万
依托单位:
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
项目状态:
已结题
起止时间:
2022 至 --
关键词:

项目摘要

项目成果

相关文献

中文摘要
翻译
Zinia是一家总部位于英国的快速发展的人工智能中小企业,拥有一个由六人组成的核心团队,为中小企业企业开发一个新的负担得起、可访问的人工智能驱动平台(Zinia),该平台可以提高决策质量,增加收入,而不需要数据科学团队。Zinia能够做出更好、更透明的决策,为客户和业务服务,从而为较小的组织提供值得信赖的、可持续的竞争优势。Zinia提供构建模型的透明度和生成决策的可解释性,并构建透明的解决方案,从无偏见的数据和决策,到有保证和证书的决策。在这个项目中,Zinia将重点关注当前的市场形势,建立动态的人工智能解决方案,在不断变化的客户情况下自学习和适应,并从稀缺和流动的数据集构建模型/决策,例如从新遇到的罕见事件数据获得的模型/决策。
英文摘要
Zinia is a rapidly growing UK-based AI SME, with a core team of six people, developing a new affordable and accessible AI-driven platform for SME businesses (Zinia) that improves decision-making quality, increases revenues without the need for data science teams.Zinia enables better, more transparent decisions that work for the customer and the business, thus enabling trustworthy, sustainable competitive advantage for smaller organizations. Zinia provides transparency of built models and explainability of generated decisions, and builds see-through solutions, ranging from bias-free data and decisions, to decisions with guarantees and certificates.With emphasis on the current market situation, within this project Zinia will build dynamic AI solutions that self-learn and adapt in changing customer situations, and build models/decisions from scarce and streaming datasets, such as those obtained from newly encountered rare-event data.
期刊论文(0)
专著(0)
科研奖励(0)
会议论文