In the last decade, the increasing amount of data produced in the financial sector has focused the attention of scholars and practitioners on their application, using Big Data Analytics techniques and Artificial Intelligence (AI) models to address interesting economic questions. An increasing number of people has experienced an exponential growth of interest in the stock market, in which assets worth billions of dollars are traded every day, with investors acting on the market with the desire to achieve a profit over their investment horizon. These analyses are increasingly used in the finance sector to deal with, but are not limited to, the following issues: (i) the actual economic problem involves lots of variables; (ii) the impact of the variables is highly nonlinear or involves interaction terms among the variables (high dimensionality of function class); and (iii) forecasting is economically more important than statistical inference. Hence, the increasing amount of data produced in the financial sector led scholars and practitioners to focus their research interest on Big Data Analytics and Artificial Intelligence techniques for the analysis of the financial domain, where accurate predictions allow investors to make informed decisions, reduce the risk of herding behavior and the bursting of financial bubbles. However, the quality of the produced insights is limited due to the high number of variables involved and their non-linear dependencies.

Furthermore, the improvements in hardware technologies (e.g. GPUs, cluster computing) and machine learning have enabled the possibility to extract relevant business value from massive and heterogeneous information (Big Data). However, deep learning-based methods have two major pitfalls: (1) the integration of stock prices, indicators and textual information (e.g. tweets, news articles) is still an unsolved problem, and (2) models are often complex and unintelligible, i.e. the "cognitive process" leading to forecasts is not transparently expressed. Indeed, the recent preponderance of machine-learning methods in financial market analysis has fostered the culture of overemphasizing model performances, rather than studying their rationale and the reasons behind their errors. However, the notion of AI models as black boxes strongly limits their acceptance in several fields, including financial forecasting. However, it is extremely important support the results with deep and meaningful insights. To this aim, we are resorting to explainable AI (XAI) methods. In fact, while producing more explainable models, it is necessary to maintain a high level of learning performance (e.g., prediction accuracy) and to enable human users to understand, appropriately trust, and effectively manage the emerging generation of artificially intelligent partners. Despite the ever-increasing availability of financial data (coming from the most disparate sources) enables us to perform analyses never conceived before, novel challenges arise from the necessity to perform elaborations on real-time streamed data and to match information coming from heterogeneous sources (multi-modality).

To address several open research issues regarding the analysis of the financial domain, this track aims at soliciting contributions highlighting challenges, state-of-the-art, and solutions to a set of currently unresolved key questions to model complex interactions and relations, and their dynamics and impact on big financial data analytics and complex financial behaviors across financial markets, regulation, and risk management.