Privacy-preserving attribute aggregation in eID federations

Filho, Walter Priesnitz; Ribeiro, Carlos; Zefferer, Thomas

doi:10.1016/j.future.2018.09.025

Cited by 13 publications

(5 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Such architecture is squarely at odds with the idea of centralized data collection and storage that undergirds the two globally dominant ecosystems that populate the Internet. Attribute-based authentication and decentralized architectures for authentication are in principle two separate issues, but, as will be argued below, the combination of attributes and decentralization gives the best guarantees for privacy protection and user empowerment (Priestnitz Filho et al, 2018).…”

Section: Sociotechnical Aspects Of Eid Systemsmentioning

confidence: 99%

Electronic identity services as sociotechnical and political-economic constructs

Dijck

Jacobs

2019

New Media & Society

View full text Add to dashboard Cite

Electronic identification services (eIDs) have become strategic services in the global governance of online societies. In this article, we argue that eIDs are sociotechnical constructs that also have political-economic dimensions. In the European context, governmental and corporate efforts to develop eIDs are shaped by legal EU frameworks, which are almost exclusively focussed on technical and legal interoperability, such as the European Interoperability Framework (EIF) and the European Interoperability Reference Architecture (EIRA). Public concerns such as privacy, security, user empowerment and control over one's personal information prompts developers to propose a decentralized, attribute-based system governed on a nonprofit, nonstate basis (DAN-eID). To illustrate our argument, we explore a single emerging eID system (IRMA; acronym for I Reveal My Attributes) that is developing in a national context (The Netherlands). We argue that developing eIDs requires more than engineering ingenuity and legal compliance; as sociotechnical and political-economic constructs, they involve negotiation of conflicting social and political values.

show abstract

Section: Sociotechnical Aspects Of Eid Systemsmentioning

confidence: 99%

Electronic identity services as sociotechnical and political-economic constructs

Dijck

Jacobs

2019

New Media & Society

View full text Add to dashboard Cite

show abstract

“…The increased prevalence of cross-border movement makes the ability to recognise legal identity even more important and identity management systems have played a key role in many strategic applications including e-governance, e-commerce, business intelligence and homeland security [8]. The focus of advanced technical research on developing open standards that address interoperability, management of privacy and identity theft prevention [10]. On the social side, there have been claims that automation of the citizen-state interface can rebuild trust although the dangers of identity theft or fraud and possible threats to the privacy of individuals resulting from use of these systems have been recognised [6].…”

Section: Conceptualising Digital Identity Systems For Refugee Managementmentioning

confidence: 99%

Reimagining Refugee Identity Systems: A Sociological Approach

Madon

Schoemaker

2019

IFIP Advances in Information and Communication Technology

View full text Add to dashboard Cite

This paper explores how the social identity of refugees shapes and is shaped through the process of registering with humanitarian organisations. Building on the recent advance of critical studies on digital identity systems for refugee management, we show how the lens of social identity is helpful in understanding the relationship between refugee information systems and refugee experiences of registration and accessing services. Identity is a key issue related to contemporary information systems yet remains an under-theorised area of investigation from a sociological perspective in the field of information systems, international development and refugee studies. Using qualitative data from 75 refugees in Bidi Bidi refugee camp in Northern Uganda, this paper showcases the centrality of refugees' social identity in determining the journey of vulnerable individuals focusing on three key dimensions. First, how the identities of refugees based on home and family in South Sudan were carried over to refugees' new location in Uganda and were later transformed through the process of registration. Second, how work and career profile of their lives in South Sudan shaped the identity of refugees, and how the absence of education credentials limited the realisation of personal aspirations. Third, how interactions between institutions and refugees are both shaped by and shape refugee identity. Our findings point to important policy implications for designing and implementing refugee identity systems.

show abstract

“…Deep learning requires large amounts of data to support training. Open image databases contain a large number of images that can be used, but due to the increasing demand for privacy protection, it is becoming increasingly difficult to collect user data to a data center for deep learning training [1,7,22,27].…”

Section: Introductionmentioning

confidence: 99%

“…7 The dataset D 𝑖 has k categories of images:𝑐 0 , 𝑐 1 ...𝑐 𝑘−1 8 𝑛𝑜𝑖𝑠𝑒𝐷𝑎𝑡𝑎 ← R (𝑀, 𝐶, 𝐻,𝑊 ) 9 𝑛𝑜𝑖𝑠𝑒𝑇 𝑎𝑟𝑔𝑒𝑡 ← randomSelect(𝑀, (𝑐 0 , 𝑐 1 ...𝑐 𝑘−1 )) 10 𝑛𝑜𝑖𝑠𝑒𝐿𝑜𝑔𝑖𝑡𝑠𝑇 𝑎𝑟𝑔𝑒𝑡 𝑗 ← 0 for j = 0,1,2,• • • , 𝑀 − 1 17 𝑛𝑜𝑖𝑠𝑒𝐷𝑎𝑡𝑎 ← 𝑛𝑜𝑖𝑠𝑒𝐷𝑎𝑡𝑎 − ∇ℓ 𝑛 (𝑛𝑜𝑖𝑠𝑒𝐷𝑎𝑡𝑎)19 gather 𝑛𝑜𝑖𝑠𝑒𝐷𝑎𝑡𝑎 and corresponding 𝑛𝑜𝑖𝑠𝑒𝐿𝑜𝑔𝑖𝑡𝑠 as 𝑛𝑜𝑖𝑠𝑒𝐷𝑎𝑡𝑎𝑠𝑒𝑡 𝑖 20 Δ𝑤 𝑡 𝑖 ← 𝑤 𝑡 − 𝑤 𝑡 𝑖 21 return (Δ𝑤 𝑡 𝑖 , 𝑛𝑜𝑖𝑠𝑒𝐷𝑎𝑡𝑎𝑠𝑒𝑡 𝑖 ) Algorithm 3: A summary of FedNKD (3) Input: local datasets D 𝑖 ; model parameters 𝑤; Output: AvgLogits for a client with model parameters 𝑤 1 GetAvgLogits(𝑤, D 𝑖 ): 2…”

mentioning

confidence: 99%

FedNKD: A Dependable Federated Learning Using Fine-tuned Random Noise and Knowledge Distillation

Zhu

Zhuang

et al. 2022

Proceedings of the 2022 International Conference on Multimedia Retrieval

View full text Add to dashboard Cite

Multimedia retrieval models need the ability to extract useful information from large-scale data for clients. As an important part of multimedia retrieval, image classification model directly affects the efficiency and effect of multimedia retrieval. We need a lot of data to train a image classification model applied to multimedia retrieval task. However, with the protection of data privacy, the data used to train the model often needs to be kept on the client side. Federated learning is proposed to use data from all clients to train one model while protecting privacy. When federated learning is applied, the distribution of data across different clients varies greatly. Disregarding this problem yields a final model with unstable performance. To enable federated learning to work dependably in the real world with complex data environments, we propose FedNKD, which utilizes knowledge distillation and random noise. The superior knowledge of each client is distilled into a central server to mitigate the instablity caused by Non-IID data. Importantly, a synthetic dataset is created by some random noise through back propagation of neural networks. The synthetic dataset will contain the abstract features of the real data. Then we will use this synthetic dataset to realize the knowledge distillation while protecting users' privacy. In our experimental scenarios, FedNKD outperforms existing representative algorithms by about 1.

show abstract

Privacy-preserving attribute aggregation in eID federations

Cited by 13 publications

References 11 publications

Electronic identity services as sociotechnical and political-economic constructs

Electronic identity services as sociotechnical and political-economic constructs

Reimagining Refugee Identity Systems: A Sociological Approach

FedNKD: A Dependable Federated Learning Using Fine-tuned Random Noise and Knowledge Distillation

Contact Info

Product

Resources

About