The rapid digitalization of the healthcare domain in recent years highlighted the need for advanced predictive methods particularly based upon deep learning methods. Deep learning methods which are capable of dealing with time- series data have recently emerged in various fields such as natural language processing, machine translation, and the Intensive Care Unit (ICU). The recent applications of deep learning in ICU have increasingly received attention, and it has shown promising results for different clinical tasks; however, there is still a need for the benchmark models as far as a handful of public datasets are available in ICU. In this thesis, a novel benchmark model of four clinical tasks on a multi-center publicly available dataset is presented; we employed deep learning models to predict clinical studies. We believe this benchmark model can facilitate and accelerate the research in ICU by allowing other researchers to build on top of it. Moreover, we investigated the effectiveness of the proposed method to predict the risk of delirium in the varying observation and prediction windows, the variable ranking is provided to ease the implementation of a screening tool for helping caregivers at the bedside. Ultimately, an attention-based interpretable neural network is proposed to predict the outcome and rank the most influential variables in the model predictions’ outcome. Our experimental findings show the effectiveness of the proposed approaches in improving the application of deep learning models in daily ICU practice.

Aquesta tesi està constituïda per diverses activitats de recerca, entre les quals trobam principalment articles i capítols publicats. Es descriu el disseny d'una joguina robòtica que facilita la metacognició dels nens. Es detallen els nous mètodes de disseny participatiu desenvolupats per donar suport a aquest procés, com també els resultats de les experiències dins l'aula que varen validar la seva eficàcia. Aquestes experiències varen fer servir l'especificació Unit of Practice d'Apple, la qual posibilita la descripció normalitzada de de les activitats pedagògiques amb els recursos d'ensenyament tecnològics. L'enfocament de la investigació llavors se centra en les especificacions de IMS, que representen recursos i activitats pedagògiques en un format que els ordinadors poden processar. S'analitzen les eines necessàries per treballar amb aquestes especificacions. Es fa una distinció entre processos de disseny en móns tancats (dirigits a grups limitats d'usuaris) i oberts (dirigits a un grup obert o universal). Es descriuen els processos participatius de disseny en ambdós casos. Es tenen especialment en consideració les eines necessàries per treballar amb l'especificació IMS LD, i es desenvolupen models conceptuals per clarificar la tipologia de les eines i els seus usos.Es proposa la participació en les comunitats de pràctica com una metodologia per avaluar les necessitats, les eines, i l'ús de l'especificació. / This thesis is composed of a number of related research activities, which are principally represented by means of published papers. The design of a robotic toy to provide support for children's meta-cognition is described. The new participatory design methods developed to support this process are detailed, together with the results of field trials which validated its effectiveness. These trials make use of Apple's Unit of Practice specification for the consistent description of pedagogic activities with technological teaching resources. The investigation of IMS' machine readable representations of pedagogy, their use, and the tooling they require, then becomes the focus for the inquiry. A distinction is drawn between closed world design processes, addressed at a circumscribed user group, and open world design addressing a wide or universal user group. Participatory design processes in both contexts are described. A particular focus is tooling for the IMS LD specification, and conceptual models are developed to clarify its tooling and use. Engagement with communities of practice is proposed as a means of addressing open world design challenges, and this methodology is used to assess user needs, tooling and use of the specification.