Human Activity Recognition (HAR) is currently recognized as a key element of a more general framework designed to perform continuous monitoring of human behaviors in the area of Ambient Assisted Living (AAL), well-being management, medical diagnosis, elderly care, rehabilitation, entertainment and surveillance in smart home environments. In this paper, an innovative HAR system, exploiting the potential of wearable devices integrated with the skills of deep learning techniques, is presented with the aim of recognizing the most common daily activities of a person at home. The designed wearable sensor embeds an Inertial Measurement Unit (IMU) and a Wi-Fi section to send data on a cloud service and to allow direct connection to the Internet through a common home router so that the user themselves could manage the installation procedure. The sensor is coupled to a CNN network designed to make inferences with the minimum possible resources to keep open the way of its implementation on low-cost or embedded devices. The system is conceived for daily activity monitor and nine different activities can be highlighted with an accuracy of 97%.

Predicting corporate bankruptcy is one of the fundamental tasks in credit risk assessment. In particular, since the 2007/2008 financial crisis, it has become a priority for most financial institutions, practitioners, and academics. The recent advancements in machine learning (ML) enabled the development of several models for bankruptcy prediction. The most challenging aspect of this task is dealing with the class imbalance due to the rarity of bankruptcy events in the real economy. Furthermore, a fair comparison in the literature is difficult to make because bankruptcy datasets are not publicly available and because studies often restrict their datasets to specific economic sectors and markets and/or time periods. In this work, we investigated the design and the application of different ML models to two different tasks related to default events: (a) estimating survival probabilities over time; (b) default prediction using time-series accounting data with different lengths. The entire dataset used for the experiments has been made available to the scientific community for further research and benchmarking purposes. The dataset pertains to 8262 different public companies listed on the American stock market between 1999 and 2018. Finally, in light of the results obtained, we critically discuss the most interesting metrics as proposed benchmarks for future studies.

A troll is usually defined as somebody who provokes and offends people to make them angry, who wants to dominate any discussion or who tries to manipulate people’s opinions. The problems caused by such persons have increased with the diffusion of social media. Therefore, on the one hand, press bodies and magazines have begun to address the issue and to write articles about the phenomenon and its related problems while, on the other hand, universities and research centres have begun to study the features characterizing trolls and to look for solutions for their identification. This survey aims at introducing the main researches dedicated to the description of trolls and to the study and experimentation of methods for their detection.

Modeling and forecasting the spread of COVID-19 remains an open problem for several reasons. One of these concerns the difficulty to model a complex system at a high resolution (fine-grained) level at which the spread can be simulated by taking into account individual features. Agent-based modeling usually needs to find an optimal trade-off between the resolution of the simulation and the population size. Indeed, modeling single individuals usually leads to simulations of smaller populations or the use of meta-populations. In this article, we propose a solution to efficiently model the Covid-19 spread in Lombardy, themost populated Italian region with about ten million people. In particular, the model described in this paper is, to the best of our knowledge, the first attempt in literature to model a large population at the single-individual level. To achieve this goal, we propose a framework that implements: i. a scale-free model of the social contacts combining a sociability rate, demographic information, and geographical assumptions; ii. a multi-agent system relying on the actor model and the High-Performance Computing technology to efficiently implement ten million concurrent agents. We simulated the epidemic scenario from January to April 2020 and from August to December 2020, modeling the government’s lockdown policies and people’s mask-wearing habits. The social modeling approach we propose could be rapidly adapted for modeling future epidemics at their early stage in scenarios where little prior knowledge is available.