ARCHANGEL - Trusted Archives of Digital Public Records

ARCHANGEL - 值得信赖的数字公共记录档案

• 批准号: EP/P03151X/1
• 负责人: John Collomosse
• 金额: $ 62.11万
• 依托单位: University of Surrey
• 依托单位国家: 英国
• 项目类别: Research Grant
• 财政年份: 2017
• 资助国家: 英国
• 起止时间: 2017 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=EP%2FP03151X%2F1
• 关键词: ARCHANGEL; Trusted; Archives; Digital; Public

The aim of ARCHANGEL is to ensure the long-term sustainability of digital archives though the design, development and trialling of transformational new distributed ledger technology (DLT) to promote accessibility and ensure integrity of content, whilst maximising its impact through novel business models for commodification and open access.Archives and Memory Institutions (AMIs) are the lens through which future generations will perceive today; they form the authoritative economic, social and cultural memory of a nation. For example, The National Archives (>15 petabytes) is one of the world's largest and oldest AMIs responsible for preserving the digital record of the UK Government e.g. key decisions made by Ministers and advice received. Some of this information is made open, some kept closed for decades. AMIs are founded upon the principle of public trust, of being neutral and completely trustworthy; the immutability and integrity of AMIs are essential to maintaining their objectivity. Yet world history is littered with examples where this objectivity has been compromised e.g. through expunging of physical records during times of political unrest. Today's digital age presents new socio-technical challenges to AMIs around safeguarding of data. Digital public records are intangible and so easy to remove or modify without that modification necessarily being detectable. Indeed in some cases records have to be modified to ensure their continued accessibility as formats change and the curation of data is also accompanied by the need to maintain associated code to render that data for presentation, often across decades. How should decisions over migration or prioritisation of maintenance be taken, or audited? What are the implications of migrations resulting in minor losses of fidelity one hundred years from now? How can the public be sure that digital content when released is fundamentally unaltered from the original? Existing archival practice is ill-equipped to respond to such issues, and is in urgent need of disruption to keep pace with our transformation into an increasingly digital society, so ensuring the integrity and impartiality of knowledge for future generations.ARCHANGEL is a 18 month socio-technical feasibility study co-creating and evaluating a novel prototype DLT service with end-users to determine how archival practices, sustainable models and public attitudes could evolve in the presence of a trusted decentralised technology to prove content integrity and ensure open access to digital public archives. From a technological standpoint, ARCHANGEL will leverage cutting-edge machine learning to collect robust digital signatures derived from digitised physical, and born-digital content, within a permissioned DLT. Both signatures and programmatic code to render content and verify its provenance and integrity will be encoded within the DLT. Novel business models for sustaining the DLT e.g. via contributed effort (proof of work) will be explored at the points of creation and consumption using a cross-AMI model in which a single DLT is contributed to by multiple AMIs, across disciplines and nations, mitigating risk of archive distortion by its operating AMI. Impact is not limited to traditional AMIs, but any digital public archive: University research data repositories (linked to DOI); better management of corporate memory in multi-nationals (e.g. financial/regulatory compliance, managing records of prior art in tech companies).To undertake this adventurous and ambitious project we have formed a strategic multi-disciplinary partnership uniting a world-leading group in multi-modal signal processing (CVSSP), the Centre for the Digital Economy (CODE) within Surrey Business School, and a consortium of AMI stakeholders including The National Archives and Tim Berners-Lee's Open Data Institute (ODI). The infrastructure will be developed with DLT platform provider Guardtime, and impact accelerated via Methods Digital.

大天使的目标是确保数字档案的长期可持续性，通过设计、开发和试验变革性的新分布式分类账技术(DLT)来促进可访问性和确保内容的完整性，同时通过商品化和开放获取的新商业模式最大限度地发挥其影响。档案和记忆机构(AMIS)是后代感知今天的透镜；它们形成了一个国家的权威经济、社会和文化记忆。例如，国家档案馆(15PB)是世界上最大和最古老的非政府组织之一，负责保存英国政府的数字记录，例如部长们做出的关键决定和收到的建议。其中一些信息是公开的，一些信息几十年来一直保持封闭。非盟驻苏特派团建立在公众信任、中立和完全值得信赖的原则基础上；非盟驻苏特派团的不变性和完整性对于保持其客观性至关重要。然而，世界历史上充斥着这种客观性受到损害的例子，例如，在政治动荡时期，通过清除实物记录。今天的数字时代给非盟驻苏特派团在保护数据方面提出了新的社会技术挑战。数字公共记录是无形的，因此很容易删除或修改，而不一定可以检测到这种修改。的确，在有些情况下，必须修改记录，以确保随着格式的变化继续获得记录，同时还需要维护相关的代码，以便经常在几十年内呈现这些数据。应该如何做出迁移或确定维护优先级的决策，或如何进行审核？从现在起一百年后，移民造成的保真度轻微损失意味着什么？公众如何确保数字内容在发布时从根本上没有改变原始内容？现有的档案实践不足以应对这些问题，迫切需要颠覆，以跟上我们向日益数字化的社会的转型步伐，从而确保未来几代人知识的完整性和公正性。ARCHANGEL是一项为期18个月的社会技术可行性研究，与最终用户共同创建和评估一项新的DLT服务原型，以确定在存在可信的分散技术的情况下，档案实践、可持续模式和公众态度如何演变，以证明内容完整性并确保开放获取数字公共档案。从技术角度来看，大天使将利用尖端机器学习在许可的DLT内收集来自数字化物理内容和天生数字内容的强大数字签名。用于呈现内容并核实其出处和完整性的签名和程序代码都将在DLT中进行编码。将在创建和消费时探索维持DLT的新商业模式，例如通过贡献的努力(工作证明)，使用跨AMI模式，其中一个DLT由跨学科和跨国家的多个AMI贡献，以减少其运营的AMI造成档案失真的风险。影响并不局限于传统的AMI，而是任何数字公共档案：大学研究数据库(链接到DOI)；更好地管理跨国企业的企业内存(例如，财务/监管合规，管理科技公司的先前技术记录)。为了承担这一大胆而雄心勃勃的项目，我们建立了一个跨学科的战略合作伙伴关系，将世界领先的多模式信号处理(CVSSP)集团、萨里商学院内的数字经济中心(CODE)以及包括国家档案馆和Tim Berners-Lee的开放数据研究所(ODI)在内的AMI利益相关者联盟联合起来。基础设施将与DLT平台提供商Guardtime一起开发，并通过方法数字加速影响。

项目成果

ARCHANGEL

大天使

• DOI: 10.1145/3209280.3229120
• 发表时间: 2018
• 作者: Collomosse J

Tamper-Proofing Video With Hierarchical Attention Autoencoder Hashing on Blockchain

• DOI: 10.1109/tmm.2020.2967640
• 发表时间: 2020-01
• 期刊: IEEE Transactions on Multimedia
• 影响因子: 7.3
• 作者: Tu Bui;Daniel Cooper;J. Collomosse;Mark Bell;A. Green;John Sheridan;J. Higgins;Arindra Das

LiveSketch: Query Perturbation for Guided Sketch-based Visual Search

LiveSketch：基于引导草图的视觉搜索的查询扰动

• 发表时间: 2019
• 作者: Collomosse J

ARCHANGEL: Tamper-Proofing Video Archives Using Temporal Content Hashes on the Blockchain

• DOI: 10.1109/cvprw.2019.00338
• 发表时间: 2019-04
• 期刊: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
• 作者: Tu Bui;Daniel Cooper;J. Collomosse;Mark Bell;Alex Green;John Sheridan;Jez Higgins;Arindra Das;Jared Keller;Olivier Thereaux;Alan W. Brown

ARCHANGEL: Trusted Archives of Digital Public Documents

ARCHANGEL：可信的数字公共文档档案

• DOI: 10.48550/arxiv.1804.08342
• 发表时间: 2018
• 作者: Collomosse J