Loading...
2 results
Search Results
Now showing 1 - 2 of 2
- On the Evaluation of Energy-Efficient Deep Learning Using Stacked Autoencoders on Mobile GPUsPublication . Falcao, Gabriel; Alexandre, Luís; Marques, J.; Frazão, Xavier; Maria, J.Over the last years, deep learning architectures have gained attention by winning important international detection and classification challenges. However, due to high levels of energy consumption, the need to use low-power devices at acceptable throughput performance is higher than ever. This paper tries to solve this problem by introducing energy efficient deep learning based on local training and using low-power mobile GPU parallel architectures, all conveniently supported by the same high-level description of the deep network. Also, it proposes to discover the maximum dimensions that a particular type of deep learning architecture—the stacked autoencoder—can support by finding the hardware limitations of a representative group of mobile GPUs and platforms.
- A Multimodal Approach to Image Sentiment AnalysisPublication . Gaspar, António; Alexandre, LuísMultimodal sentiment analysis is a process for the classi- cation of the content of composite comments in social media at the sentiment level that takes into consideration not just the textual content but also the accompanying images. A composite comment is normally represented by the union of text and image. Multimodal sentiment analysis has a great dependency on text to obtain its classi cation, because image analysis can be very subjective according to the context where the image is inserted. In this paper we propose a method that reduces the text analysis dependency on this kind of classi cation giving more importance to the image content. Our method is divided into three main parts: a text analysis method that was adapted to the task, an image classi er tuned with the dataset that we use, and a method that analyses the class content of an image and checks the probability that it belongs to one of the possible classes. Finally a weighted sum takes the results of these methods into account to classify content according to its sentiment class. We improved the accuracy on the dataset used by more than 9%.