Coupling Learning in Big Data

大数据中的耦合学习

• 批准号: DP150102442
• 负责人: Prof Longbing Cao
• 金额: --
• 依托单位: University of Technology Sydney
• 依托单位国家: 澳大利亚
• 项目类别: Discovery Projects
• 财政年份: 2015
• 资助国家: 澳大利亚
• 起止时间: 2015-01-01 至 2018-12-31
• 项目状态: 已结题
• 来源: https://dataportal.arc.gov.au/RGS/Web/Grants/DP150102442
• 关键词: Coupling; Learning; Big; Data

Big data features complex coupling relationships within and between diverse entities in various forms and layers. This fundamentally challenges existing learning theories, which usually assume that data is independent and identically distributed (IID). This indicates that such IID tools may either be inapplicable for big data or capture an incomplete or even biased picture of the ground truth in big data. Hence, this project aims to invent breakthrough theories and effective tools for systematically modelling and learning sophisticated couplings embedded in big data applications. The outcomes are expected to enhance Australia's leading role in data science research and lift data intelligence-driven productivity and economic growth in a changing world.

大数据具有各种形式和层次的各种实体内部和之间的复杂耦合关系。这从根本上挑战了现有的学习理论，这些理论通常假设数据是独立同分布的（IID）。这表明，这些IID工具可能不适用于大数据，或者捕获大数据中基础事实的不完整甚至有偏见的图片。因此，该项目旨在发明突破性的理论和有效的工具，用于系统地建模和学习嵌入大数据应用程序中的复杂耦合。预计这些成果将增强澳大利亚在数据科学研究方面的主导作用，并在不断变化的世界中提升数据智能驱动的生产力和经济增长。