Alexandros Ntoulas’ Topics
1.TITLE: Fake Content Detection
ABSTRACT (up to 100 words): Fake news and information is abundant in the World Wide Web as people are using this medium for anything from political to personal and monetary gains. Recently Facebook was under scrutiny for allowing fake information to propagate through their Web services. Detecting fake news is of paramount importance as it can affect public opinion and steer it towards specific directions.In this thesis we will implement a service that will be able to detect inaccurate or untrue content and classify it accordingly.
We will investigate and implement deep learning techniques to ensure accurate detection and we will experiment with different variations to improve accuracy.
Required skills: programming (Java/Python/HTML+Javascript/Git), databases, machine learning
2.TITLE: Data Storage for Internet of Things (IoT) Data
ABSTRACT (up to 100 words): In typicall IoT applications we observe that there is a great volume of data generated from sensors that have the following common characteristics: a) they are small in size (payload) b) they are frequent enough, and c) they are typically repeated. Most of the smart devices today upload their data to the cloud where it is later being processed. However, due to the characteristics of the data, we are currently wasting a lot of bandwidth.
In this thesis we will study and implement data compression and transmission techniques in order to minimize the wasted bandwidth and storage of the data in the cloud.
Required skills: programming (Java/Python/HTML+Javascript/Git), databases, machine learning
NAME & POSITION OF THE SUPERVISOR: Alexandros Ntoulas, Assistant Professor
LAB/GROUP, DEPARTMENT, INSTITUTION where the thesis will be executed: Department of Informatics, National and Kapodistrian University of Athens