Search

The growth of social media worldwide has brought social benefits and challenges. One problem we highlight is the proliferation of hate speech on social media. We propose a novel method for detecting hate speech in texts using Cross-Lingual Learning. Our approach uses transfer learning from Pre-Trained Language Models (PTLM) with large corpora available to solve problems in languages with fewer resources for the specific task. The proposed methodology comprises four stages: corpora acquisition, the PTLM definition, training strategies, and evaluation. We carried out experiments using Pre-Trained Language Models in English, Italian, and Portuguese (BERT and XLM-R) to verify which best suited the proposed method. We used corpora in English (WH) and Italian (Evalita 2018) as the source language and the OffComBr-2 corpus in Portuguese (the target language). The results of the experiments showed that the proposed methodology is promising: for the OffComBr-2 corpus, the best state-of-the-art result was obtained (F1-measure = 92%). • The development of a new methodology for hate speech detection. • Portuguese hate speech detection using Cross-Lingual Learning. • Up to 20% performance improvement over other models using the OffComBr-2 corpus. [ABSTRACT FROM AUTHOR]

Thanks to the built-in GPS device embedded in almost all smartphones, the facility of tracking users’ positions fostered new research opportunities. Among these opportunities, of particular interest in this work is the field of route and destination prediction. Suggesting a user to take a deviation to avoid a congested route is among the potential benefits of our research. Many of the approaches available in the literature consolidate the Markov model as suitable to prediction. Moreover, the Prediction by Partial Matching (PPM) compression technique has presented encouraging results for predicting route and destination. Thus, this paper proposes a novel predictor that combines Markov model with PPM technique, extracting the better of these two approaches. Our user-personalized predictor is able to predict the route and destination automatically in a real-time manner, including places never visited by the user. We evaluated our model with real world data collected from 21 users, obtaining a precision rate between 63% and 82%. [ABSTRACT FROM AUTHOR]

This article proposes a semi-automatic technique for the annotation of people in photographs. The technique uses context and content information and is based on a weighted sum of estimators, which results in a list of the person’s contacts that are more likely to be present in a photograph. Machine learning methods, such as multivariable linear regression and slope function, are adopted to filter and weight the estimators and eigenfaces for face recognition. The article also describes the results of experiments that were performed with a collection of 4050 photographs with 365 different people, which indicate that the proposed technique outperforms techniques that adopt only context or only content using as a performance metric the H-Hit rate of correct annotations. [ABSTRACT FROM AUTHOR]