Healthcare as a sector, with all the longitudinal data it collects around patients across their life cycles, is now in a place where it can take advantage of what Artificial Intelligence (AI), data science, and in specific machine Learning (ML) have to victims. Endless opportunities such as diagnostics, and prognostics, to care personalization and improving healthcare quality and efficiency are a few examples of such offers.
A key prerequisite for precision healthcare is the estimation of disease progression from the current patient state. Patient trajectories as a means of illustrating the temporal disease progression and disease correlations are considered as one of the best methods to characterize patients and diseases. The trajectories also have a prognostic/predictive potential where preceding steps can be used as a basis for predicting the most likely next step in disease progression, leading to precise treatment. However, lack of inscrutability of the AI/ML-based techniques has hampered pervasive AI in healthcare systems. Graph-based methods are one solution for that respect.
In this talk, we will introduce graph-based approaches for collecting, storing, visualizing, analyzing, and modeling healthcare data, and discuss how they outperform the other available approaches. We will also discuss challenges around health data. Access to such a valuable source of healthcare data holds great promise for policy evaluation and large-scale biomedical research investigations. However, access to healthcare data is a big challenge, since the introduction of the European General Data Protection Regulation (GDPR) and the Patient Data Act in Sweden. As a consequence of these legislations, the ability to access large samples of "real" patient data for researchers, pharmaceutical companies, and other related stakeholders is limited. There is also a need to deal with challenges around health data including biased, unbalanced,
First attempt would be to share the "real" data while maintaining patient privacy. Various computational and AI approaches have been pursued to prevent identity and attribute disclosures through different anonymization techniques or data perturbations. But providing raw data, at any degree of specificity, leads to a tradeoff between privacy and data utility.
With the availability of powerful methods, generating synthetic health data is now considered as an alternative that maintains data utility and fidelity while preserving patient privacy.
Artificiell intelligens och maskininlärning
Förinspelat + På plats
Forskare (även studerande)
Erik Brandt Föreläsare
Chief Data Scientist
Hallandia V AB
Data scientist with a passion for statistical analysis, machine learning and graph algorithms. Currently chief data scientist at SHAARPEC by Hallandia V AB.
Kobra Etminani Föreläsare
Associate Professor, Docent
Center for Applied Intelligent Systems Research (CAISR) in Health, Halmstad University
Kobra (Farzaneh) Etminani is an associate Professor, Docent, working at the Center for Applied Intelligent Systems Research (CAISR), Halmstad University, Sweden. She is Deputy Profile Manager for CAISR Health, a Swedish funded research profile. She is also part of the extended management group for Information Driven Care research program (IDC) at Halmstad University.
She manages the Real-World Evidence (RWE) research projects together with Health Data Center (HDC), that includes several research projects together with Region Halland (a regional Swedish healthcare system), analytics companies, and big Pharma.
She has worked on various topics and application areas within Machine Learning (ML), Artificial Intelligence (AI), Data Mining, and Deep Learning (DL) in the last decade, focused on healthcare, district heating, and mobility. Her main research interest is focused on solving real-world problems, which is focused on a healthier society, with the help of AI and ML, if possible and applicable. Her recent research focus is on patient trajectories and eXplainable AI (XAI) in precision healthcare