Jorge Morais

Abstract
Climate change affects global temperature and precipitation patterns. These effects, in turn, influence the intensity and, in some cases, the frequency of extreme environmental events, such as forest fires, hurricanes, heat waves, floods, droughts, and storms. In general, these events can be particularly conducive to the appearance of plant pests and diseases. The availability of models and a data collection system is crucial to manage pests and diseases in sustainable agricultural ecosystems. Agricultural ecosystems are known to be complex, multivariable, and unpredictable. It is important to anticipate crop pests and diseases in order to improve its control in a more ecological and economical way (e.g., precision in the use of pesticides). The development of an intelligent monitoring and management platform for the prevention of pests and diseases in olive groves at Trás-os- Montes region will be very beneficial. This platform must: a) integrate data from multiple data sources such as sensory data (e.g., temperature), biological observations (e.g., insect counts), georeferenced data (e.g., altitude) or digital images (e.g., plant images); b) systematize these data into a regional repository; c) provide relevant forecasts for pest and diseases. Convolutional Neural Networks (CNNs) can be a valuable tool for the identification and classification of images acquired by Internet of Things (IoT).

Abstract

Abstract
Recommender systems form a class of Artificial Intelligence systems that aim to recommend relevant items to the users. Due to their utility, it has gained attention in several applications domains and is high demanded for research. In order to obtain successful models in the recommendation problem in non-prohibitive computational time, different heuristics, architectures and information filtering techniques are studied with different datasets. More recently, machine learning, especially through the use of deep learning, has driven growth and expanded the sequential recommender systems development. This research focuses on models for managing sequential recommendation supported by session-based recommendation. This paper presents the characterization in the specific theme and the state-of-the-art towards study object of the thesis: the adaptive recommendation to mitigate the information overload in online environments.

Abstract

Abstract
The evacuation of buildings in case of fire is a sensitive issue for civil society that also motivates the academic community to develop and study solutions to improve the efficiency of evacuating these spaces. The study of human behavior in fire emergencies has been one of the areas that have deserved the attention of researchers. However, this modeling of human behavior is difficult and complex because it depends on factors that are difficult to know and that vary from country to country. In this paper, a paradigm shift is proposed which, instead of focusing on modeling the behavior of occupants, focuses on conditioning this behavior by providing real-time information on the most efficient evacuation routes. Making this information available to occupants is possible with a solution that takes advantage of the growing use of the IoT (Internet of Things) in buildings to help occupants adapt to the environment. Supported by the IoT, multi-agent recommender systems can help users to adapt to the environment and provide the occupants with the most efficient evacuation routes. This paradigm shift is achieved through a context-based multi-agent recommender system based on contextual data obtained from IoT devices, which recommends the most efficient evacuation routes at any given time. The obtained results suggest that the proposed solution can improve the efficiency of evacuating buildings in the event of a fire; for a scenario with two hundred people following the system recommendations, the time they take to reach a safe place decreases by 17.7%.