Named Entity Recognition (NER) is a challenging learning task of identifying and classifying entity mentions in texts into predefined categories. In recent years, deep learning (DL) methods empowered by distributed representations, such as word- and character-level embeddings, have been employed in NER systems. However, for information extraction in Police narrative reports, the performance of a DL-based NER approach is limited due to the presence of fine-grained ambiguous entities. For example, given the narrative report “Anna stole Ada’s car”, imagine that we intend to identify the VICTIM and the ROBBER, two sub-labels of PERSON. Traditional NER systems have limited performance in categorizing entity labels arranged in a hierarchical structure. Furthermore, it is unfeasible to obtain information from knowledge bases to give a disambiguated meaning between the entity mentions and the actual labels. This information must be extracted directly from the context dependencies. In this paper, we deal with the Hierarchical Entity-Label Disambiguation problem in Police reports without the use of knowledge bases. To tackle such a problem, we present HELD, an ensemble model that combines two components for NER: a BLSTM-CRF architecture and a NER tool. Experiments conducted on a real Police reports dataset show that HELD significantly outperforms baseline approaches.

Humans need to communicate. Out of this basic need combined with the Web, a vast amount of text has been generated on a daily basis. Given the presence of a lot of information allocated in different resources, it becomes vital to enable machines to understand spoken and written texts. This chapter presents how Deep Learning techniques can solve Natural Language Processing (NLP) tasks (e.g., Text Classification and Sentence Summarization), aiming to benefit from the computational power currently available and the low need for feature engineering when using these models. Initially, some essential concepts about NLP and Deep Learning are presented. Then, different pre-processing and textual representation techniques are explained to be used as input in Deep Learning models. Finally, it is shown how to apply the knowledge acquired in real applications of NLP.

Deep Learning is a subarea of Machine Learning that uses neural networks with successive layers of data representation, which allow the performance of more complex tasks. Generally, these models produce better results when working with large volumes of data. However, in some situations, obtaining labeled data in large quantities for training these networks is not feasible. The meta-learning strategy aims to mitigate this problem, enabling learning models to learn quickly from other models initially trained for different tasks. This work introduces some meta-learning techniques, focusing on their use with Deep Learning models to solve tasks with fewer data. ResumoAprendizado Profundo é uma subárea de Aprendizado de Máquina que utiliza redes neurais com sucessíveis camadas de representação dos dados, as quais permitem a realização de tarefas mais complexas. Em geral, esses modelos produzem melhores resultados quando trabalhados com grandes volumes de dados. Entretanto, em algumas situações, não é possível obter dados rotulados em grande quantidade para o treinamento dessas redes. A estratégia de meta-learning visa mitigar esse problema, fazendo com que modelos de aprendizagem consigam aprender, de forma rápida, a partir de outros modelos inicialmente treinados para diferentes tarefas. Este trabalho introduz algumas técnicas de meta-learning, focando em seu uso com modelos de Aprendizado Profundo para a resolução de tarefas com quantidade de dados reduzida.

Este artigo descreve as atividades desenvolvidas pela ação “TI por Elas”, que é organizada pelo Centro Acadêmico de Ciência da Computação (CACC) em parceria com o Programa de Educação Tutorial/Conexões de Saberes (PET - TI) do Campus da UFC em Quixadá. Essa ação tem como objetivo principal cultivar o interesse de mulheres pelaárea da Tecnologia da Informação (TI), tanto como disseminar o que é a computação e sua importância. Tendo em vista o público da área, a presente ação também pretende ajudar a diminuir a evasão das mulheres que já se encontram no âmbito da TI.