Browsing by Author "Kallio, Johanna"
Now showing 1 - 2 of 2
Results Per Page
Sort Options
- Challenges of learning human digital twin: case study of mental wellbeing: Using sensor data and machine learning to create HDTPublication . Vildjiounaite, Elena; Kallio, Johanna; Kantorovitch, Julia; Kinnula, Atte; Ferreira, Simão; Rodrigues, Matilde; Rocha, NunoHuman Digital Twin (HDT) is a powerful tool to create a virtual replica of a human, to be used for example for designing interactions with physical systems, preventing cognitive overload, managing human capital, and maintaining a healthy and motivated workforce. Building human twins is a challenging task due to the need to reliably represent each corresponding human being, and the fact that human beings notably differ from each other. Therefore, relying solely on expert knowledge is insufficient, and human twins must learn the specifics of each individual in order to accurately represent them. This paper focuses on AI methods for modelling the mental wellbeing of knowledge workers because the mounting cognitive demands of both white-collar and blue-collar work lead to employees’ stress, and stress leads to diminished creativity and motivation, increased sick leaves, and in severe cases, accidents, burnouts, and disabilities. This paper describes the main building blocks of AI-based detectors of mental stress and highlights the main challenges and future directions of research., which are expected to be relevant also for HDT learning in other domains because the high degree of individuality is ubiquitous in all human activities.
- From controlled to chaotic: Disparities in laboratory vs real-world stress detectionPublication . Ferreira, Simão; Rodrigues, Fátima; Kallio, Johanna; Coelho, Filipe; Kyllonen, Vesa; Rocha, Nuno; Rodrigues, Matilde A.; Vildjiounaite, Elena; Ferreira, Simão; Rodrigues, MatildeThis paper explores the discrepancies between laboratory and real-world stress detection, emphasizing the pronounced differences in data loss, data preprocessing, feature design, and classifier selection. Laboratory studies offer a controlled environment that optimizes data quality, whereas real-world settings introduce chaotic and unpredictable elements, coupled with a diverse range of human behaviours, resulting in substantial data loss and compromised data quality. We discuss the development of stress detectors for two distinct types of data: physiological and behavioural. We also address the specific challenges associated with designing effective stress detection systems for each data type and compare the features and classifiers used in both laboratory and real-world contexts. Additionally, this paper proposes future research directions aimed at crafting stress detectors that are robust and effective in real-life scenarios.