Federated learning is the process of developing machine learning models over datasets distributed across data centers such as hospitals, clinical research labs, and mobile devices while preventing ...data leakage. This survey examines previous research and studies on federated learning in the healthcare sector across a range of use cases and applications. Our survey shows what challenges, methods, and applications a practitioner should be aware of in the topic of federated learning. This paper aims to lay out existing research and list the possibilities of federated learning for healthcare industries.
Federated Machine Learning Yang, Qiang; Liu, Yang; Chen, Tianjian ...
ACM transactions on intelligent systems and technology,
03/2019, Volume:
10, Issue:
2
Journal Article
Peer reviewed
Today’s artificial intelligence still faces two major challenges. One is that, in most industries, data exists in the form of isolated islands. The other is the strengthening of data privacy and ...security. We propose a possible solution to these challenges: secure federated learning. Beyond the federated-learning framework first proposed by Google in 2016, we introduce a comprehensive secure federated-learning framework, which includes horizontal federated learning, vertical federated learning, and federated transfer learning. We provide definitions, architectures, and applications for the federated-learning framework, and provide a comprehensive survey of existing works on this subject. In addition, we propose building data networks among organizations based on federated mechanisms as an effective solution to allowing knowledge to be shared without compromising user privacy.
Während materielle Schadensersatzansprüche für Datenschutzverletzungen in der Praxis eine untergeordnete Rolle zu spielen scheinen und verhältnismäßig einfach festzustellen und zu beziffern sind, ...bereitet die in Art. 82 DSGVO vorgesehene Ersatzfähigkeit immaterieller Schäden den Gerichten Kopfzerbrechen. Eine richtungsweise Entscheidung zu immateriellen Schadensersatzansprüchen für DSGVO-Verletzungen fällte der EuGH Anfang Mai 2023 in der Rechtssache C‑300/21. Es ist das erste Urteil aus einer langen Reihe an Vorabentscheidungsersuchen zur Auslegung des Art. 82 DSGVO. Nach wie vor interpretationsbedürftig bleibt jedoch, wie ein immaterieller Schaden nun konkret festzustellen und zu bemessen ist. Nach einer kurzen Zusammenfassung der Kernaussagen des EuGH befasst sich dieser Beitrag daher mit diesem praxisrelevanten Problem und möchte – insbesondere unter Berücksichtigung etablierter Instrumentarien der deutschen und österreichischen Rechtspraxis – Lösungswege für die mitgliedstaatlichen Gerichte aufzeigen.
Emerging as a buzzword, the General Data Protection Regulation (GDPR) has had immense implications on global data protection regimes. The GDPR appears as a worldwide standard for protecting personal ...data based on the omnibus legal substance, extensive extraterritorial scope, and influential market of the European Union (EU). It resulted in a global wave where countries are either adopting new legislation or modifying existing data privacy laws to comply with the GDPR. Historically, the South Asian region, abode to one-fifth of the world’s people, has strong trade and economic ties with Europe. As reflected in current bilateral or multilateral trade agreements, the EU tends to be one of the largest trading partners of most South Asian countries. Therefore, it is understandable that the EU’s norms, laws, policies, particularly the GDPR, would have far-reaching impacts on South Asian countries. However, the issue has not been yet evaluated in legal academic settings that require an analysis of GDPR’s overview and its impacts on South Asian privacy regimes. The findings of this doctrinal legal study, together with the sharing of a brief overview of the GDPR and South Asian privacy regimes, reiterate the influence of GDPR in this region. The findings of this research also have the prospects to enlighten the stakeholders in understanding the GDPR and its implications on global as well as South Asian privacy regimes. This article concludes with several suggestions and policy alternatives that policymakers can explore in South Asia and beyond in designing their potential personal data protection policy strategies.
La tecnologia blockchain rappresenta un fenomeno rivoluzionario per diversi settori della vita dell’economia e non solo. Anzi, da un lato la necessità di adattare i tradizionali settori al ...cambiamento tecnologico, dall’altro le possibilità che tale tecnologia fornisce in termini di efficienza e velocizzazione, hanno portato la blockchain a essere presa sempre più in considerazione. La questione ancora in via di definizione, tuttavia, riguarda il trattamento dei dati, in un ecosistema trasparente e accessibile. L’obiettivo del presente scritto è, dunque, quello di analizzare l’impatto e le applicazioni di tale tecnologia in tutti i settori in qui essa è potenzialmente applicabile e, di conseguenza, le implicazioni in termini di privacy. Blockchain technology represents a revolutionary phenomenon for various sectors of life in the economy and beyond. Indeed, on the one hand the need to adapt traditional sectors to technological change, and on the other the possibilities this technology provides in terms of efficiency and speed, have led blockchain to be increasingly taken into consideration. The issue still to be defined, however, concerns the processing of data, in a transparent and accessible ecosystem. The aim of this paper is, therefore, to analyse the impact and applications of this technology in all the sectors where it is potentially applicable and, consequently, the implications in terms of privacy
As World Economic Forum’s definition of personal data as ‘the new “oil” – a valuable resource of the 21st century’ shows, large-scale data processing is increasingly considered the defining feature ...of contemporary economy and society. Commercial and governmental discourse on data frequently argues its benefits, and so legitimates its continuous and large-scale extraction and processing as the starting point for developments in specific industries, and potentially as the basis for societies as a whole. Against the background of the General Data Protection Regulation, this article unravels how general discourse on data covers over the social practices enabling collection of data, through the analysis of high-profile business reports and case studies of health and education sectors. We show how conceptualisation of data as having a natural basis in the everyday world protects data collection from ethical questioning while endorsing the use and free flow of data within corporate control, at the expense of its potentially negative impacts on personal autonomy and human freedom.
This study represents the first research effort to explore the transition from traditional teaching into distance teaching in Swedish schools enforced by covid-19. Governments made gradual and ...injudicious decisions to impede the spread of the pandemic (covid-19) in 2020. The enactment of new measures affected critical societal functions and included travel restrictions, closing of borders, school closures and lockdowns of entire countries worldwide. Social distancing became the new reality for many, and for many teachers and students, the school closure prompted a rapid transition from traditional to distance education. This study aims to capture the early stages of that transition. We distributed a questionnaire to teachers' (n = 153) to gain insights into teacher and school preparedness, plans to deliver distance education, and teachers' experience when making this transition. Results show that the school preparedness was mainly related to technical aspects, and that teachers lack pedagogical strategies needed in the emerging learning landscape of distance education. Findings reveal four distinct pedagogical activities central for distance education in a crisis, and many challenges faced during the transition. While preparedness to ensure continuity of education was halting, schools and teachers worked with tremendous effort to overcome the challenges. Results expand on previous findings on school closure during virus outbreaks and may in the short-term support teachers and school leaders in making informed decisions during the shift into distance education. The study may also inform the development of preparedness plans for schools, and offers a historical documentation.
Can the act of assigning a score to someone constitute a decision? This, in essence, is the question the Court of Justice of the European Union (CJEU) had to answer in Case C-634/21. And the Court’s ...answer is yes, following in the footsteps of the Advocate General’s opinion on the case. Rendered on 7 December, this ruling was eagerly awaited as it was the first time the Court had the opportunity to interpret the notorious Article 22 of the General Data Protection Regulation (GDPR) prohibiting decisions “based solely on automated processing".
Person re-identification is a key challenge for surveillance across multiple sensors. Prompted by the advent of powerful deep learning models for visual recognition, and inexpensive RGB-D cameras and ...sensor-rich mobile robotic platforms, e.g. self-driving vehicles, we investigate the relatively unexplored problem of cross-modal re-identification of persons between RGB (color) and depth images. The considerable divergence in data distributions across different sensor modalities introduces additional challenges to the typical difficulties like distinct viewpoints, occlusions, and pose and illumination variation. While some work has investigated re-identification across RGB and infrared, we take inspiration from successes in transfer learning from RGB to depth in object detection tasks. Our main contribution is a novel method for cross-modal distillation for robust person re-identification, which learns a shared feature representation space of person’s appearance in both RGB and depth images. In addition, we propose a cross-modal attention mechanism where the gating signal from one modality can dynamically activate the most discriminant CNN filters of the other modality. The proposed distillation method is compared to conventional and deep learning approaches proposed for other cross-domain re-identification tasks. Results obtained on the public BIWI and RobotPKU datasets indicate that the proposed method can significantly outperform the state-of-the-art approaches by up to 16.1% in mean Average Precision (mAP), demonstrating the benefit of the distillation paradigm. The experimental results also indicate that using cross-modal attention allows to improve recognition accuracy considerably with respect to the proposed distillation method and relevant state-of-the-art approaches.11Code: https://github.com/frhf/cross-modal-distillation-reidentification.
•Cross-modal distillation training procedure to transfer embedding representation•Exploiting the intrinsic relation between depth and RGB.•Ideal deep feature distillation needs to take place from depth to RGB.•Novel cross-modal gated attention mechanism improves our distillation approach.
PDF
