论文标题
区块链作为在智能环境中转移学习的推动者
Blockchain as an Enabler for Transfer Learning in Smart Environments
论文作者
论文摘要
在智能系统的机器学习模型中体现的知识通常与耗时且昂贵的过程相关联，例如大规模数据收集，数据标记，网络培训和模型的微调。在部署在不同环境中的智能系统之间共享和重复使用这些详尽的模型（称为转移学习）将有助于为用户提供服务，并加速在智能建筑和智能城市应用等环境中摄入的智能系统。在这种情况下，支持AI的环境之间的沟通和知识交流取决于复杂的系统，系统系统，数字资产系统及其依赖链的网络，这些网络几乎不遵循传统信息系统的集中式架构。相反，它需要一个自适应分散的系统体系结构，该系统体系结构由数据出处，工作流透明度和过程参与者验证等功能授权。在这项研究中，我们提出了一个基于区块链和知识图技术的分散和自适应软件框架，该框架以透明且可信赖的方式支持知识交换和IOT支持环境之间的互操作性。
The knowledge, embodied in machine learning models for intelligent systems, is commonly associated with time-consuming and costly processes such as large-scale data collection, data labelling, network training, and fine-tuning of models. Sharing and reuse of these elaborated models between intelligent systems deployed in a different environment, which is known as transfer learning, would facilitate the adoption of services for the users and accelerates the uptake of intelligent systems in environments such as smart building and smart city applications. In this context, the communication and knowledge exchange between AI-enabled environments depend on a complicated networks of systems, system of systems, digital assets, and their chain of dependencies that hardly follows the centralized schema of traditional information systems. Rather, it requires an adaptive decentralized system architecture that is empowered by features such as data provenance, workflow transparency, and validation of process participants. In this research, we propose a decentralized and adaptive software framework based on blockchain and knowledge graph technologies that supports the knowledge exchange and interoperability between IoT-enabled environments, in a transparent and trustworthy way.
