National Edge AI Hub for Real Data: Edge Intelligence for Cyber-disturbances and Data Quality

用于真实数据的国家边缘人工智能中心：针对网络干扰和数据质量的边缘智能

• 批准号: EP/Y028813/1
• 负责人: Rajiv Ranjan
• 金额: $ 1309.6万
• 依托单位: Newcastle University
• 依托单位国家: 英国
• 项目类别: Research Grant
• 财政年份: 2024
• 资助国家: 英国
• 起止时间: 2024 至 无数据
• 项目状态: 未结题
• 来源: https://gtr.ukri.org/projects?ref=EP%2FY028813%2F1
• 关键词: National; Edge; AI; Hub; Real

The Hub will address two challenges introduced by the use of Edge Computing (EC) to support emerging AI algorithms: dealing with cyber disturbances and managing data quality. The Hub will achieve these through a unique 3x3x3x2 matrix that reflects the complexity of these systems:(1) 3 real-world application domains(2) 3 tiers of EC architecture(3) 3 ground-breaking research work streams(4) 2 industry engagement work streamsTheir inter-relationships will be examined by a multi-disciplinary team with track records in EC architecture (Newcastle, Cardiff, St. Andrews, UWS, Imperial, Hull), foundational AI and Data Quality (Southampton, Durham, QUB, Swansea), wireless communication (Cardiff, UWS), device malfunction, attack detection and prevention (Newcastle, Lancaster, Cardiff, Warwick), and AI security (Lancaster, Swansea, Durham, Warwick). This network will enable us to engage with regional development agencies in these areas.Applications include autonomous electric vehicles, energy security and remote healthcare. At the Newcastle Urban Observatory test-bed, a world-leading UK-funded effort collaborating with sensor system manufacturers, software companies and others, we will use interactions across 3 tiers of EC architecture: sensors (Tier 1), edge devices that control them (Tier 2), and cloud-based data storage and processing (Tier 3), to identify the benefit of these interactions in the real-world data processing. The agenda will be underpinned by activities in 5 interrelated work streams.We have strategies in:Embedding Equality, Diversity & Inclusion: We are committed to EDI policies of UKRI and EPSRC Councils and EDI policies of our members. From these, we will form our guiding principle around EDI for members to adhere to in all matters related to the Hub, including recruitment, research, workshops, project allocations, outreach activities, etc. All core committees will have an EDI champion. We will ensure that activities are fair, free from bias and preference of any kind, and uphold the respect and integrity of all members. The Hub is constituted of members from diverse ethnic backgrounds, races, and gender and has intrinsically diverse and multicultural characteristics. We will actively encourage students from under-represented groups to pursue industry-funded PhDs with the Hub. The PDRA requirement in the Hub will, while maintaining the best talent, offer equal opportunity to candidates of all backgrounds, disabilities, sexual orientations, gender, and ethnicity. The Hub will use institutional infrastructure to support the well-being of staff and members.Intellectual Property Management: While most research outcomes will be made public (e.g., software open access), some may be subject to patents. Participating universities have commercialisation offices to identify, assess, protect, manage and commercially develop IP to maximise national benefits from public investment in research, which we will use to commercialise significant outcomes. Information about services, standards used, and other technical details will be made public to attract industrial partners and to promote training in the new technologies. Non-technical press releases and notes will be available to general audiences. An in-principle agreement has been reached with consortium members that each shall retain ownership of any background IP contributed to the project and that the ownership of project-generated IP shall be shared based on respective partner contributions. Hub activities will in general follow the National Principles of Intellectual Property Management for Publicly Funded Research and this will be applied to each project managed under the feasibility fund. Non-Disclosure Agreements with commercial partners will be in place to manage sensitive information. Specific terms regarding IP will be further defined in a collaborative research agreement before the commencement of the project.

该中心将解决使用边缘计算（EC）来支持新兴AI算法所带来的两个挑战：处理网络干扰和管理数据质量。该中心将通过一个独特的3x3x3x2矩阵来实现这些目标，该矩阵反映了这些系统的复杂性：（1）3个实际应用领域（2）3层电子商务架构（3）3个突破性的研究工作流程（4）2个行业参与工作流程它们之间的相互关系将由一个在电子商务架构方面有良好记录的多学科团队进行研究（纽卡斯尔、卡迪夫、圣安德鲁斯、UWS、帝国、船体）、基础人工智能和数据质量（南安普顿、达勒姆、QUB、斯旺西），无线通信（卡迪夫，UWS），设备故障，攻击检测和预防（纽卡斯尔、兰开斯特、卡迪夫、沃里克），以及AI安全（兰开斯特、斯旺西、达勒姆、沃里克）。该网络将使我们能够与这些领域的区域发展机构合作。应用包括自动驾驶电动汽车，能源安全和远程医疗保健。在纽卡斯尔城市观测台测试台，一个世界领先的英国资助的努力与传感器系统制造商，软件公司和其他人合作，我们将使用跨3层EC架构的交互：传感器（第1层），控制它们的边缘设备（第2层），以及基于云的数据存储和处理（第3层），以确定这些交互在现实世界的数据处理中的好处。我们在以下方面制定了战略：嵌入平等、多样性和包容性：我们致力于UKRI和EPSRC理事会的EDI政策以及我们成员的EDI政策。在此基础上，我们将围绕EDI形成我们的指导原则，供成员在与中心有关的所有事务中坚持，包括招聘，研究，研讨会，项目分配，外展活动等。我们将确保活动是公平的，没有任何形式的偏见和偏好，并维护所有成员的尊重和诚信。该中心由来自不同民族背景、种族和性别的成员组成，具有内在的多样性和多元文化特征。我们将积极鼓励来自代表性不足的群体的学生在该中心攻读行业资助的博士学位。该中心的PDRA要求将在保持最佳人才的同时，为所有背景，残疾，性取向，性别和种族的候选人提供平等的机会。该中心将利用机构基础设施来支持工作人员和成员的福祉。知识产权管理：虽然大多数研究成果将被公开（例如，软件开放获取），有些可能会受到专利。参与的大学设有商业化办公室，以识别，评估，保护，管理和商业开发知识产权，以最大限度地提高国家从公共研究投资中获得的利益，我们将利用这些投资将重大成果商业化。有关服务、使用的标准和其他技术细节的信息将公布于众，以吸引工业合作伙伴，并促进新技术的培训。非技术性新闻稿和说明将提供给一般读者。已与联合体成员达成原则性协议，每个成员应保留对项目贡献的任何背景知识产权的所有权，项目产生的知识产权的所有权应根据各自合作伙伴的贡献分享。中心的活动一般将遵循公共资助研究的知识产权管理国家原则，这将适用于可行性基金管理的每个项目。与商业合作伙伴的保密协议将到位，以管理敏感信息。有关知识产权的具体条款将在项目开始前的合作研究协议中进一步界定。