Advanced vector architectures for future applications

Although there has been a lot of progress in the domain of automatic knowledge discovery (e.g., deep learning), we believe humans play a central role in the data analysis task. Therefore, we propose to build a visual stream processing system for supporting the analysis and exploration of data streams in real-time, by exploiting human’s natural ability for discovering patterns. Our work combines stream processing and data storage techniques [2, 3], with data visualization theory. We study strategies for visualizing different types of data considering constraints related to real time and data volume. This paper presents our stream processing and visualization system adapted to the requirements of the neuroscience domain