We have experienced a rapidly increasing growth of data and information. A proliferation of automated systems, small scale computing devices, sensor networks, and data capture technologies has contributed to the production of large volumes of data. Data set growth outpaces available storage capacity. The focus of data processing and analysis has moved from offline batch processing of data to the incremental handling of online data streams. Online data streams may originate from sources such as mobile devices, autonomous vehicles, industrial monitoring, brain-computer interfaces, financial and meteorological systems, health care, stock market, web traffic and clickstreams, to name a few. Their prominence in real-world systems, along with the necessity of modeling, analyzing, and understanding these systems, has brought new challenges, greater demands, and new research directions. 

Data stream modeling is fundamentally based on incremental learning methods that process data continuously as an attempt to find similarities in spatio-temporal features and, thereafter, provide insights about the phenomenon that governs the data. Data streams are characterized by nonstationarity, nonlinearity, and heterogeneity; they are potentially endless and may be subject to changes of various kinds. Direct application of computational intelligence and learning algorithms to data streams is very often infeasible because it is difficult to maintain all the data in memory. Particular challenges faced in stream modeling concerns how to handle the inherent uncertainty in the data, and how the resulting model can be understood by non-experts. 

The key research questions addressed in this Special Session are (i) how to obtain accurate and explainable models from uncertain data streams; and (ii) how to exploit uncertainty and vague reasoning to better explain adaptive models at any time step. 

The Special Session aims at bringing together theorists and practitioners who apply lifelong learning methods for sequential (and uncertain) data analysis to exchange and discuss ideas that enrich traditional approaches, e.g., computational methods for static datasets. The special session gets together experts from different research communities including (but not limited to):

The OLUD special session intends to facilitate interdisciplinary discussion on recent advancements in state-of-the-art online learning and pattern recognition methods as well as their use in applied domains.