Care data
Care data
Care data
Research Data Platform
Research Data Platform
Research Data Platform
Im rasch voranschreitenden Gesundheitswesen ermöglicht die Integration künstlicher Intelligenz (KI) eine Verbesserung der Patientenergebnisse. Um die Forschung in diesem Bereich zu erleichtern, werden hochwertige und robuste Daten benötigt, die die Entwicklung von KI-Modellen ermöglichen. Da Gesundheitsdaten sensibel sind und geschützt werden müssen, ist das Teilen von Daten unter Wahrung der Patientenrechte ein wichtiges Thema. Die Bereitstellung qualitativer Daten unter Wahrung der Privatsphäre ist daher für Forscher von unschätzbarem Wert. Aus diesem Grund haben wir uns zum Ziel gesetzt, hochgradig authentische, synthetische Daten zu generieren und sie Forschern auf einer öffentlichen, dezentralen Plattform zur Verfügung zu stellen, um die Forschung zu KI in der Pflege voranzutreiben.
Im rasch voranschreitenden Gesundheitswesen ermöglicht die Integration künstlicher Intelligenz (KI) eine Verbesserung der Patientenergebnisse. Um die Forschung in diesem Bereich zu erleichtern, werden hochwertige und robuste Daten benötigt, die die Entwicklung von KI-Modellen ermöglichen. Da Gesundheitsdaten sensibel sind und geschützt werden müssen, ist das Teilen von Daten unter Wahrung der Patientenrechte ein wichtiges Thema. Die Bereitstellung qualitativer Daten unter Wahrung der Privatsphäre ist daher für Forscher von unschätzbarem Wert. Aus diesem Grund haben wir uns zum Ziel gesetzt, hochgradig authentische, synthetische Daten zu generieren und sie Forschern auf einer öffentlichen, dezentralen Plattform zur Verfügung zu stellen, um die Forschung zu KI in der Pflege voranzutreiben.
Im rasch voranschreitenden Gesundheitswesen ermöglicht die Integration künstlicher Intelligenz (KI) eine Verbesserung der Patientenergebnisse. Um die Forschung in diesem Bereich zu erleichtern, werden hochwertige und robuste Daten benötigt, die die Entwicklung von KI-Modellen ermöglichen. Da Gesundheitsdaten sensibel sind und geschützt werden müssen, ist das Teilen von Daten unter Wahrung der Patientenrechte ein wichtiges Thema. Die Bereitstellung qualitativer Daten unter Wahrung der Privatsphäre ist daher für Forscher von unschätzbarem Wert. Aus diesem Grund haben wir uns zum Ziel gesetzt, hochgradig authentische, synthetische Daten zu generieren und sie Forschern auf einer öffentlichen, dezentralen Plattform zur Verfügung zu stellen, um die Forschung zu KI in der Pflege voranzutreiben.
Our Approach - Data Architecture & Privacy
Our Approach - Data Architecture & Privacy
Our Approach - Data Architecture & Privacy
Our approach to data architecture and data protection involves a four-step process that includes data extraction [1], data quality and analysis [2], as well as decentralized data repository [3], and finally, machine learning [4].
Our approach to data architecture and data protection involves a four-step process that includes data extraction [1], data quality and analysis [2], as well as decentralized data repository [3], and finally, machine learning [4].
Our approach to data architecture and data protection involves a four-step process that includes data extraction [1], data quality and analysis [2], as well as decentralized data repository [3], and finally, machine learning [4].
Using privacy-preserving decentralized deep learning approaches, predictive models for fall prognosis are to be developed that take into account all relevant risk factors. The data used for this purpose comes from two large nursing institutions and a startup. To protect privacy during data processing, generative deep learning models are to be trained in the institutions that can generate synthetic but as authentic as possible patient data.
Using privacy-preserving decentralized deep learning approaches, predictive models for fall prognosis are to be developed that take into account all relevant risk factors. The data used for this purpose comes from two large nursing institutions and a startup. To protect privacy during data processing, generative deep learning models are to be trained in the institutions that can generate synthetic but as authentic as possible patient data.
Using privacy-preserving decentralized deep learning approaches, predictive models for fall prognosis are to be developed that take into account all relevant risk factors. The data used for this purpose comes from two large nursing institutions and a startup. To protect privacy during data processing, generative deep learning models are to be trained in the institutions that can generate synthetic but as authentic as possible patient data.
By embedding them in standardized runtime environments, realized through container technologies, the models can be shared among facilities as well as with external AI researchers. This allows for realistic data to be provided without having to share real patient data. Using the data integration and analysis methods developed in the project, as well as the AI application, alternative relevant questions, data, and outcomes could also be evaluated.
The standardized runtime environment should enable fast development of AI applications and data analysis, both externally and on site without integration effort. This innovative infrastructure enables for the first time the development and validation of guideline-compliant AI-based nursing fall prevention across multiple facilities.
By embedding them in standardized runtime environments, realized through container technologies, the models can be shared among facilities as well as with external AI researchers. This allows for realistic data to be provided without having to share real patient data. Using the data integration and analysis methods developed in the project, as well as the AI application, alternative relevant questions, data, and outcomes could also be evaluated.
The standardized runtime environment should enable fast development of AI applications and data analysis, both externally and on site without integration effort. This innovative infrastructure enables for the first time the development and validation of guideline-compliant AI-based nursing fall prevention across multiple facilities.
By embedding them in standardized runtime environments, realized through container technologies, the models can be shared among facilities as well as with external AI researchers. This allows for realistic data to be provided without having to share real patient data. Using the data integration and analysis methods developed in the project, as well as the AI application, alternative relevant questions, data, and outcomes could also be evaluated.
The standardized runtime environment should enable fast development of AI applications and data analysis, both externally and on site without integration effort. This innovative infrastructure enables for the first time the development and validation of guideline-compliant AI-based nursing fall prevention across multiple facilities.