Abstract

Abstract
The importance of early detection of diseases with high-mortality is crucial to save lives. Deep Learning algorithms are recurrently used by many researchers that aim to model the progression and treatment of these conditions. There is growing evidence that the complexity of a Deep Learning model is correlated to its performance: the deeper the network, the more accurate it is. However, as the topology deepens, training gets more demanding: (1) increased need of data, (2) increased computational costs, and (3) increased time for evaluation, fine-tuning, and subsequent feedback-based activities inherent to Data Science, with direct impact on the exploration towards finding the best model, due to an inherent trial-and-error approach. We hypothesize that there exist (domain-specific) architectural patterns that, if applied during the model exploration phase, allow an overall improvement of the training performance. Should it be true, it would significantly reduce the exploration phase length, contributing to both Medicine and Computer Science fields.

Abstract
Determining subgraph frequencies is at the core of several graph mining methodologies such as discovering network motifs or computing graphlet degree distributions. Current state-of-the-art algorithms for this task either take advantage of common patterns emerging on the networks or target a set of specific subgraphs for which analytical calculations are feasible. Here, we propose a novel network generic framework revolving around a new data-structure, a Condensed Graph, that combines both the aforementioned approaches, but generalized to support any subgraph topology and size. Furthermore, our methodology can use as a baseline any enumeration based census algorithm, speeding up its computation. We target simple topologies that allow us to skip several redundant and heavy computational steps using combinatorics. We were are able to achieve substantial improvements, with evidence of exponential speedup for our best cases, where these patterns represent up to 97% of the network, from a broad set of real and synthetic networks.

Abstract
Poor dietary behaviours are commonly associated with severe chronic diseases such as cardiovascular diseases, diabetes and obesity. Personalized food recommendation systems can be an important motivation to stimulate and inform people on best dietary practices by suggesting healthy foods and nutritionally balanced meals adjusted to their preferences and daily routines. The development of such systems require the process and integration of data available from different sources with different representations. FILLET is an intelligent platform for nutrition capable of collecting and integrating data from multiple sources including recipe websites, food blogs and nutrition databases. Components were developed for web scraping, identifying ingredients, estimating nutritional content and matching ingredients with food products from retailers to support a meal recommendation and shopping list assistance services. We present for each component the challenges identified in the literature and the ones we faced in their development, describing our approach and the lessons learned that can contribute to the future improvement of the platform and the development of related platforms.

Abstract
We used BPMN diagrams to identify indicators that can assist teachers in their intervention actions to support students' self-regulation and co-regulation in an asynchronous e-learning context. The use of BPMN modeling, by making explicit the tasks and procedures implicit in the intervention of the e-learning teacher, also exposed which data were available for developing decision-support indicators, as well as the relevant moments for carrying out interventions. Such indicators can help e-learning teachers focus their interventions to support self-regulation and co-regulation of learning, as well as enabling the creation of live data dashboards to support decision-making for those interventions, thus this process can contribute to devise better instruments for teacher intervention in support of self-regulation and co-regulation of student learning.

Abstract