An open lecture by:
Thanos Petsakos
When: Friday, 11 February, 18:00-21:00
Where: Online via ZOOM
Organized by: The Graduate Master Class Series, School of Graduate and Professional Education
Lecture Description
The discussion will focus on research, which culminated in a paper published in the ACM, and an award (“Best Presenter”) at the 2021 DSIT Conference held in Shanghai in June 2021. The team (comprised of 4 authors) addressed the problem of classifying documents into categories for the insurance market. One of the most arduous tasks in any business is to sort through piles of documents. Often this is done using copious amounts of time. The most basic, and time-consuming task is the visual inspection of each document in order to discover its type. The work performed by the team, aimed to address the problem of examining each document submitted by the customer of a large Insurance firm, to discover whether it is the one that was requested. The aforementioned problem was addressed with supervised machine learning methods. In particular two approaches were followed; in the first approach documents were handled as images and convolutional neural networks (CNN) were employed. In the second approach, text was extracted from the images with the use of Optical Character Recognition (OCR) algorithms and classification was performed (aided by Natural Language Processing techniques) based on the contents of each document. The entire process, together with conclusions, limitations and future work will be presented.
Thanos Petsakos
Thanos Petsakos is a data analyst and data scientist, currently working as a Senior Technology Consultant | Data and Analytics at EY Greece. He has almost 15 years’ experience in the Finance and Banking sector where he held the positions of Business Analyst, Senior MIS analyst and Data Scientist. Up until Dec 2021 he held the position of Head of MIS and Business Analytics in Eurobank’s Private Banking Business. Thanos also works as a Data Science Instructor at the Big Blue Data Academy, and also volunteers as an Analytics tutor at CuriousIQ. He has also worked as a research assistant. He has aided a professor of Finance and Quantitative Methods (at ALBA) and a professor of Applied Econometrics (at the University of Athens) with various data-driven projects. Thanos holds a Bachelor’s degree (BSc) in Business and Economics, a Master’s (MSc) in Finance (hons), an MBA (focus in Banking & Financial Services) (hons) and a Master’s (MSc) in Data Science (hons). He is currently in the process of receiving his MSc in Artificial Intelligence at the University of Bath in the U.K. Among his professional certifications are a Joint Certificate in SAS Business Analytics and Data Mining, and a certification in Financial Modeling & Valuation (CFI FMVA – pending). Thanos has participated in several projects (both as a student and as a professional). He has co-authored an article in Kathimerini Newspaper (based on research conducted as part of his MBA) as well as a paper (on Machine Learning – Document Classification), which has been published by the Association of Computing Machinery (ACM) in 2021.