Desenvolvendo Modelos de Deep Learning para Aplicações Multimídia no Tensorflow

Abstract

The availability of massive quantities of data, combined with increasing computational capabilities, makes it possible to develop more precise Machine Learning algorithms. These new tools provide advances in areas such as Natural Language Processing and Computer Vision, allowing efficient processing of images, text and audio. Now, cognitive functionalities, such as learning, recognition and detection, can be used in multimedia applications to create mechanisms beyond traditional capture, streaming and presenta- tion uses. Methods based on Deep Learning became state-of-the-art in several Multi- media challenges. This short course presents the grounds and ways to develop models using Deep Learning. It prepares the participant to: (1) understand and develop mod- els based on Deep Neural Networks, Convolutional Neural Networks (CNN), Recurrent Neural Networks, including LSTM and GRU; (2) apply the Deep Learning models to solve problems within the multimedia domain like Image Classification, Facial Recogni- tion, Object Detection, Video Scenes Classification. The Python programming language is shown alongside TensorFlow, a package for developing Deep Learning models.