Research Seminar by Gary ANG Meng Kiat

Talk #1: Investment and Risk Management with Online News and Heterogeneous Networks

Abstract: Stock price movements in financial markets are influenced by large volumes of news from diverse sources on the web, e.g., online news outlets, blogs, social media. Extracting useful information from online news for financial tasks, e.g., forecasting stock returns or risks, is however challenging due to the low signal-to-noise ratios of such online information. Assessing the relevance of each news article to the price movements of individual stocks is also difficult, even for human experts.

In this talk, we will present our research in the paper, “Investment and Risk Management with Online News and Heterogeneous Networks”, which has been recently published in the ACM Transactions on the Web. In this paper, we propose the Guided Global-Local Attention-based Multimodal Heterogeneous Network (GLAM) model, which comprises novel attention-based mechanisms for multimodal sequential and graph encoding, a guided learning strategy, and a multitask training objective. GLAM uses multimodal information, heterogeneous relationships between companies and leverages significant local responses of individual stock prices to online news to extract useful information from diverse global online news relevant to individual stocks for multiple forecasting tasks. Our experiments with multiple datasets show that GLAM out-performs other state-of-the-art models on multiple forecasting tasks, and important investment and risk management applications.

Talk #2: Learning Dynamic Multimodal Networks

Abstract: Capturing and modeling relationship networks between entities or objects, and attributes associated with such networks is an important challenge in network or graph learning. In this talk, we will present our research on models that capture and model dynamic multimodal networks.  Examples of such networks include: networks of design objects in user interfaces (UI), e.g., links between UI screens and their constituent UI elements, where the design objects may be associated with multimodal attributes such as visual UI screen and element images, textual UI descriptions and code; and dynamic networks between companies, e.g., commercial relationships between companies that evolve across time, where the companies may be associated with dynamic multimodal attributes such as time-series of numerical stock prices, textual news, and categorical event attributes. In this talk, we will present an overview of works focusing on different aspects of learning dynamic multimodal networks, that have been accepted by conferences/journals such as the Annual Meeting of the Association for Computational Linguistics, ACM Transactions on Interactive Intelligent Systems, IEEE International Conference on Big Data.

Gary Ang is a PHD candidate in the School of Computing and Information Systems, Singapore Management University, supervised by Professor Lim Ee Peng. His research interests include network and time-series modelling in domains such as user interfaces, finance, and sustainability.