idw – Informationsdienst Wissenschaft
Nachrichten, Termine, Experten
Nachrichten, Termine, Experten
idw - Informationsdienst
Wissenschaft
This press-release presents the progress of work on the integration of PAROMA_MED platform and showcases the same via a collaborative workshop to gather citizen´s perspective on the solution built by PAROMA_MED consortium.
PAROMA-MED is a 36 months project, funded by EU. It is integrating the concepts of federated machine learning, AI, Data Spaces and privacy preserving data federation, privacy aware communications networks into prototypes, which can be demonstrated and help in accelerating the development of healthcare solutions that unleash the power and the value of, otherwise underexploited, healthcare data. It believes in implementing sustainable practices and contributes to the health vertical. The goal is to develop Sustainable technology by design and build trustworthiness for its uptake. With collaborative efforts we wish to empower health industry, individuals and related communities with the knowledge and tools needed to navigate the digital landscape.
PAROMA-MED platform builds on:
Federated machine learning approach: that allows multiple devices or servers to train a model collaboratively without sharing raw data. Instead of sending data to a central server, each device (or node) trains a local model using its own data and then sends only model updates (such as gradients or weights) to a central server, which aggregates them to improve the global model. For privacy- and security- assured services and applications in federative cross-border environments technologies, a hybrid- cloud framework is developed which is focusing on an elevated role of the edge nodes as the entities that ensure adequate protection and strict adherence to security and privacy rules. PAROMA-MED builds on top of the federation principles foreseen by Data Spaces and establishes an alternative federation mode that is based on zero data movement. According to this concept specific parts of the Data Space protocol workflows and information models have been adapted to cater for code to data principles.
Traceability and Integrity: Integration of watermarking technology in the PAROMA-MED project to ensure the traceability of data and AI models. This technology allows the detection of the origin of leaks if data or models have been compromised in a malicious way. Additionally, when data needs to be shared between edge nodes or when providing access to remote physicians for annotation purposes, the watermarking can also control the integrity of the data.
Path computation: demonstration of application that relates to a communication network, and path computation in such a network based on different weights assigned to meet the network security requirements with minimum latency and includes the innovative Path Computation Engine (Patent filed) conceived in the context of the PAROMA-MED project Network and Interconnect Platform.
Use Case: The PAROMA-MED use case focuses on heart segmentation in medical image analysis, highlighting AI's role in healthcare. Traditional centralized learning has security risks due to sensitive data transmission. With federated learning approach, we are transmitting models instead of data. Models are trained locally in hospitals, then aggregated centrally to enhance performance. The PAROMA-MED platform processes CT scans within hospitals to compute myocardial thickness, a key diagnostic parameter.
Recently the project PAROMA-MED showcased its platform and demonstrated its solutions in a collaborative workshop with the project TRANSCEND. The joint workshop organized by PAROMA-MED and TRANSCEND aimed to educate participants on these critical issues and gather feedback on their use of the platform. A dynamic communication environment was created through a combination of expert input and interactive discussions. With the help of the Toolbox developed in the TRANSCEND project (see here), it was possible to identify citizens' perspectives on the PAROMA-MED platform (see here). The participants could get hands on the platform and could administer hypothetical consent options regarding their medical records. Afterwards they had the opportunity to see how their options are being enforced and how this is perceived by a potential data scientist with particular scientific interest in their data. Overall, PAROMA-MED achieved to make a very comprehensive demonstration of what “functional GDRP” means. This has achieved to elevate the awareness of citizens regarding both their fundamental rights and the fact that it is feasible to have fine grained control over their data and the ways these may be used. This engaging workshop brought insights and experiences, into a comprehensive examination of topics as transparency, privacy and modes of use of technical solutions. The participants departed from the workshop with an enhanced comprehension of their digital rights and the importance of their personal data. Additionally, concrete recommendations were formulated to further develop the PAROMA-MED platform. The citizens' perspectives also flow directly into the technology development process and the suggestions are being processed by the PAROMA-MED consortium.
The joint workshop by PAROMA-MED and TRANSCEND represented a significant contribution to the learning and capturing the experience of stakeholders and their perspective will be included in the refining of the platform.
PAROMA-MED continues to engage with stakeholders and exchange with our innovative tools designed to empower citizens and enhance digital autonomy.
k.koutsopoulos@qualtek.eu
bassi@eurescom.eu
https://paroma-med.eu/
https://transcend-project.eu/empowering-digital-citizens-insights-from-the-parom...
PAROMA-MED
Merkmale dieser Pressemitteilung:
Journalisten
Informationstechnik
überregional
Forschungsprojekte, wissenschaftliche Weiterbildung
Englisch
Sie können Suchbegriffe mit und, oder und / oder nicht verknüpfen, z. B. Philo nicht logie.
Verknüpfungen können Sie mit Klammern voneinander trennen, z. B. (Philo nicht logie) oder (Psycho und logie).
Zusammenhängende Worte werden als Wortgruppe gesucht, wenn Sie sie in Anführungsstriche setzen, z. B. „Bundesrepublik Deutschland“.
Die Erweiterte Suche können Sie auch nutzen, ohne Suchbegriffe einzugeben. Sie orientiert sich dann an den Kriterien, die Sie ausgewählt haben (z. B. nach dem Land oder dem Sachgebiet).
Haben Sie in einer Kategorie kein Kriterium ausgewählt, wird die gesamte Kategorie durchsucht (z.B. alle Sachgebiete oder alle Länder).
