Applying deep learning on metadata data as a competitive accelerator

While data quality is an important issue in view of the developments in the area of data science. Reaping the benefits of artificial intelligence is impossible in the presence of data quality problems. Many data science projects need to start with a data cleaning phase, which at times can be very costly. A variety of data cleaning techniques exist, each targeted at different types of data quality problems. As manual data cleaning is costly, and even not feasible in the presence of very large amounts of data, algorithmic data cleaning is an interesting area of research. The specific goal of the project is to perform research on the use of analytics, deep learning and reinforcement learning for the creation, improvement and management of meta-data, with the aim to improve the quality of data and metadata so that new products and services can be developed faster, and regulatory constraints can be met more easily. The results we aim for,  are novel and intelligent machine learning algorithms for the identification of data quality problems of different types in tabular data.