One of the most persistent problems in information management is semantic interoperability. This is true for classical business computing as well as for open data on the Internet. Solving this problem is often prerequisite for tackling other downstream tasks.
The classical task in information processing is providing the relevant information to users. A key element for success is the ability to integrated preferences of users into the models employed.
A significant part of information is contained in content of documents. Extracting structured information from unstructured documents is a critical ability in automating many information processing tasks.
Increasing amounts of information are being produced by unreliable and unknown actors. Beyond understanding the meaning of data and documents, it is also critical to provide the capabilities to assess their quality and reliability.