Many real-world projects are impacted by uncertainties that can appear from many different sources. This project should investigate the range of machine learning techniques for dealing with the uncertainty problem and look particularly at probabilistic and machine reasoning approaches.We would encourage students to showcase a demo based on test data and then experiment with algorithms to produce a solution that meets the above problem.

Expected outcome:

• Document or report detailing the different approaches and their pros and cons.
• Working demo showing the results on top of data e.g. probabilities of each of the outcomes with respect to uncertainty.
•  A report on the algorithms and techniques used.

Prerequisites:

STEM field (Science, Technology, Engineering, Mathematics).

As SAP enterprise software continuously deals with structured data stored in tables, CVS files or Excel, we would like to investigate methods of deriving a knowledge graph from structured data. The challenge requires the students to know the concept and use of knowledge graphs as a prerequisite, and then to first research and finally explore options that can achieve the above. Students will be guided by the SAP team which is working on real-world customer problems that require the use of knowledge graphs for the purpose of data modeling.

Expected outcome:

• Select 3-5 approaches from academia for inferring user intent on the basis of UI interactions and justify reason for selection.
• Apply the approaches to UI interaction data sets and evaluate suitability/fit of the approach to the data type, inference quality, limitations, and other learnings.
• Final deliverable: report documenting the chosen approaches and evaluation results including pros & cons per approach, comparison/contrast, and ranking of approaches.

Prerequisites

STEM field, Semantic technology fundamentals