São Paulo Federal Institute of Education, Science and Technology

About: São Paulo Federal Institute of Education, Science and Technology is a education organization based out in São Paulo, Brazil. It is known for research contribution in the topics: Context (language use) & Computer science. The organization has 1707 authors who have published 2374 publications receiving 11333 citations.

Exposing Computer Generated Images by Eye’s Region Classification via Transfer Learning of VGG19 CNN

Abstract: The advance of computer graphics techniques comes revolutionizing games and movie’s industries. Creating very realistic characters totally from computer graphics models is, nowadays, a reality. However, this advance comes with a big price: the realism of images is so big that it is difficult to realize when we are facing a computer generated image or a real photo. In this paper we propose a new approach for highly realistic computer generated images detection by exploring inconsistencies into the region of the eyes. Such inconsistencies are captured exploring the expression power of features extracted via transfer learning approach with VGG19 Deep Neural Network model. Unlike the state-of-the-art approaches, which looks to evaluate the entire image, proposed method focuses in specific regions (eyes) where computer graphics modeling still needs improvements. Experiments conducted over two different datasets containing extremely realistic images achieved an accuracy of 0.80 and an AUC of 0.88.

Reviewing the world's edible mushroom species: A new evidence-based classification system

Abstract: Wild mushrooms are a vital source of income and nutrition for many poor communities and of value to recreational foragers. Literature relating to the edibility of mushroom species continues to expand, driven by an increasing demand for wild mushrooms, a wider interest in foraging, and the study of traditional foods. Although numerous case reports have been published on edible mushrooms, doubt and confusion persist regarding which species are safe and suitable to consume. Case reports often differ, and the evidence supporting the stated properties of mushrooms can be incomplete or ambiguous. The need for greater clarity on edible species is further underlined by increases in mushroom-related poisonings. We propose a system for categorizing mushroom species and assigning a final edibility status. Using this system, we reviewed 2,786 mushroom species from 99 countries, accessing 9,783 case reports, from over 1,100 sources. We identified 2,189 edible species, of which 2,006 can be consumed safely, and a further 183 species which required some form of pretreatment prior to safe consumption or were associated with allergic reactions by some. We identified 471 species of uncertain edibility because of missing or incomplete evidence of consumption, and 76 unconfirmed species because of unresolved, differing opinions on edibility and toxicity. This is the most comprehensive list of edible mushrooms available to date, demonstrating the huge number of mushrooms species consumed. Our review highlights the need for further information on uncertain and clash species, and the need to present evidence in a clear, unambiguous, and consistent manner.

Length of Stay at Multiple Destinations of Tourism Trips in Brazil

Abstract: This study applied a shared heterogeneity duration model to tourists’ length of stay at different locations of multidestination trips. This analysis helps to understand tourists’ behaviors and to predict their length of stay according to relevant variables. Such information can be applied to the development of efficient marketing strategies aiming to push the average length of stay to the desired direction, and to develop “on the fly” service provision and revenue management strategies. The focus on multiple destination trips offers an innovative analytical perspective. A large data set of 309,000 visits to Brazilian destinations was analyzed. Several empirical findings regarding determinants of tourists’ length of stay were obtained. Positively skewed distributions for duration and hazard functions were found to best fit observed data. Shared heterogeneity was found to statistically improve the explanatory capacity of duration models when multidestination tourism trips data are analyzed.

Explainable Matrix - Visualization for Global and Local Interpretability of Random Forest Classification Ensembles

Abstract: Over the past decades, classification models have proven to be essential machine learning tools given their potential and applicability in various domains. In these years, the north of the majority of the researchers had been to improve quantitative metrics, notwithstanding the lack of information about models' decisions such metrics convey. This paradigm has recently shifted, and strategies beyond tables and numbers to assist in interpreting models' decisions are increasing in importance. Part of this trend, visualization techniques have been extensively used to support classification models' interpretability, with a significant focus on rule-based models. Despite the advances, the existing approaches present limitations in terms of visual scalability, and the visualization of large and complex models, such as the ones produced by the Random Forest (RF) technique, remains a challenge. In this paper, we propose Explainable Matrix (ExMatrix), a novel visualization method for RF interpretability that can handle models with massive quantities of rules. It employs a simple yet powerful matrix-like visual metaphor, where rows are rules, columns are features, and cells are rules predicates, enabling the analysis of entire models and auditing classification results. ExMatrix applicability is confirmed via different examples, showing how it can be used in practice to promote RF models interpretability.