The MSc in Data Science is a master program that strives to connect academic knowledge and business practice.
In the context of the Demokritos-QUALCO Fellowship programme, 5 students of the academic year 2021-2022, have been awarded master's thesis scholarships and started working in October 2022.
Important Info
- The master's thesis must be completed within 6 months. This is a requirement to obtain the scholarship.
- QUALCO will supervise closely the students along with the supervisors on the part of the master program.
- Each master's thesis scholarship amounts to 3.000 €.
Topics
Below you will find the five proposed subjects by QUALCO, along with a small description.
| Proposed subject | Small description | |
|---|---|---|
| 1 | Text analytics approaches to multichannel information summarisation on Fintech customers | Leverage the available free-text sources of information (e.g. digital documents, emails, text messages, call centre agent comments) on customers of Fintech companies (e.g. debt servicers) to derive summarised, 360-degree views of customer history and behaviour. |
| 2 | Enhancement of text-based treatment outcome modelling in the credit sector using state-of-the-art mechanisms | Straightforward, unsophisticated text-based modelling techniques have shown some promise in predictive modelling of treatment outcomes for non-performing credit. The objective of this exercise is to use more advanced techniques, including e.g. attention mechanisms, to investigate improvements in model accuracy and generalisability. |
| 3 | Text-driven data exploration and report generation | Use past reporting deliverables, metadata on the structure of available data, and NLP techniques to develop a plain text-driven exploration interface for credit portfolio data. |
| 4 | AI-driven early warnings of defaults on performing credit | Use AI techniques and data on performing loans falling into default to identify early warning signals and build predictive models allowing preemptive treatment of high-risk customers. |
| 5 | AI-drive rehabilitation indicators for non-performing credit | Use AI techniques and data on non-performing loans and their treatment to identify loans that are likely to proceed to a mutually acceptable solution (e.g. payment arrangement), whether in the short term (e.g. offer acceptance), or in the longer term (e.g. successful resolution of the debt). |
About the Demokritos–QUALCO Fellowship programme
Education is one of the strategic pillars of the Institute of Informatics and Telecommunications, NCSR Demokritos (IIT) and the institute's vision is to bring research and industry closer; to connect science with several industrial domains. One of the initiatives undertaken towards this direction is the Demokritos-QUALCO Fellowship programme.
IIT and QUALCO launched this 5-year collaboration in October 2021 and it aims to support research and development (R&D) projects, which include PhD and Post-Doctoral research, scholarships to MSc students, scientific projects, innovative workshops and competitions. Among the goals of the program is to help MSc students explore the FinTech (financial technology) landscape and ecosystem. Furthermore, it supports young PhD and Post doc researchers to delve into their research in the Fintech field using AI and Big Data.
