Abstract
Nowadays, there is the paradox of technology: although smartphones have revolutionized our way of living, bringing convenience and connectivity, they have also introduced new challenges, notably distracted driving. This paper addresses the issue of visual distraction, one of the main contributors to traffic accidents, through the development of an innovative system that combines the application of convolutional neural networks and the functionality of mobile devices. The adopted methodology focused on the collection of a broad set of images to train an artificial intelligence model capable of classifying a qualitative variable with two distinct categories: attention and distraction of a driver. In particular, the study concentrated on creating a mobile application that uses a smartphone's camera to monitor the driver and issue auditory alerts if it detects prolonged distraction. The achieved results highlighted the efficacy of the model, especially after its optimization for the TensorFlow Lite format, suitable for implementation on mobile devices due to its efficiency in terms of speed and resource consumption.

Abstract

Abstract
Recent studies have shown that the microbiome can impact cancer development, progression, and response to therapies suggesting microbiome-based approaches for cancer characterization. As cancer-related signatures are complex and implicate many taxa, their discovery often requires Machine Learning approaches. This review discusses Machine Learning methods for cancer characterization from microbiome data. It focuses on the implications of choices undertaken during sample collection, feature selection and pre-processing. It also discusses ML model selection, guiding how to choose an ML model, and model validation. Finally, it enumerates current limitations and how these may be surpassed. Proposed methods, often based on Random Forests, show promising results, however insufficient for widespread clinical usage. Studies often report conflicting results mainly due to ML models with poor generalizability. We expect that evaluating models with expanded, hold-out datasets, removing technical artifacts, exploring representations of the microbiome other than taxonomical profiles, leveraging advances in deep learning, and developing ML models better adapted to the characteristics of microbiome data will improve the performance and generalizability of models and enable their usage in the clinic.

Abstract
Vertical pulp and paper production is challenging from a process point of view. Managers must deal with floating bottlenecks, intermediate storage levels, and by-product production to control the whole process while reducing unexpected downtimes. Thus, this paper aims to address the integrated lot sizing and scheduling problem considering continuous digester production, multiple paper machines, and a chemical recovery line to treat by-products. The aim is to minimize the total production cost to meet customer demands, considering all productive resources and encouraging steam production (which can be used in power generation). Production planning should define the sizes of production lots, the sequence of paper types produced in each machine, and the digester working speed throughout the planning horizon. Furthermore, it should indicate the rate of byproduct treatment at each stage of the recovery line and ensure the minimum and maximum storage limits. Due to the difficulty of exactly solving the mixed integer programming model representing this problem for realworld instances, mainly with planning horizons of over two weeks, constructive and improvement heuristics are proposed in this work. Different heuristic combinations are tested on hundreds of instances generated from data collected from the industry. Comparisons are made with a commercial Mixed-Integer and Linear Programming solver and a hybrid metaheuristic. The results show that combining the greedy constructive heuristic with the new variation of a fix-and-optimize improvement method delivers the best performance in both solution quality and computational time and effectively solves realistic size problems in practice. The proposed method achieved 69.41% of the best solutions for the generated set and 55.40% and 64.00% for the literature set for 1 and 2 machines, respectively, compared with the best solution method from the literature and a commercial solver.

Abstract
Autonomous Mobile Robots (AMR) are seeing an increased introduction in distinct areas of daily life. Recently, their use has expanded to intralogistics, where forklift type AMR are applied in many situations handling pallets and loading/unloading them into trucks. One of the these vehicles requirements, is that they are able to correctly identify the location and status of pallets, so that the forklifts AMR can insert the forks in the right place. Recently, some commercial sensors have appeared in the market for this purpose. Given these considerations, this paper presents a comparison of the performance of two different approaches for pallet detection: using a commercial off-the-shelf (COTS) sensor and a custom developed application based on Artificial Intelligence algorithms applied to an RGB-D camera, where both the RGB and depth data are used to estimate the position of the pallet pockets.

Abstract
Lot-sizing and scheduling in a job shop environment is a fundamental problem that appears in many industrial settings. The problem is very complex, and solutions are often needed fast. Although many solution methods have been proposed, with increasingly better results, their computational times are not suitable for decision-makers who want solutions instantly. Therefore, we propose a novel greedy heuristic to efficiently generate production plans and schedules of good quality. The main innovation of our approach represents the incorporation of a simulation-based technique, which directly generates schedules while simultaneously determining lot sizes. By utilising priority rules, this unique feature enables us to address the complexity of job shop scheduling environments and ensures the feasibility of the resulting schedules. Using a selection of well-known rules from the literature, experiments on a variety of shop configurations and complexities showed that the proposed heuristic is able to obtain solutions with an average gap to Cplex of 4.12%. To further improve the proposed heuristic, a cooperative coevolutionary genetic programming-based hyper-heuristic has been developed. The average gap to Cplex was reduced up to 1.92%. These solutions are generated in a small fraction of a second, regardless of the size of the instance.