2. Research and Development of Deep Learning Technologies
The purpose of WP2 is the research and development of deep learning methods for portfolio management with the aim of optimizing specific performance indicators, such as PnL (Profit and Loss) and Sharpe Ratio. Also, the development of deep learning methods for natural language understanding, enabling semantic analysis and sentiment extraction from text streams originating from heterogeneous media. Finally, the development of multimodal information fusion methods (time series, sentiment indicators, etc.) for financial portfolio management. The tasks corresponding to WP 2 are the following:
Task 2.1: Development of deep learning methods for portfolio management
Task 2.2: Development of deep learning methods to extract sentiment from social media streams
Task 2.3: Development of multimodal information fusion methods for portfolio management