On this page, you will find some thesis opportunities (I'll try to keep this list updated).. Feel free to contact me if you want to pursue your own ideas.
This thesis proposal focuses on the application of reinforcement learning (RL) techniques to enable autonomous flight of drone swarms in beyond-visual-line-of-sight (BVLOS) scenarios. The objective is to devise and implement an algorithm that allows a swarm of drones to efficiently accomplish designated tasks while adhering to BVLOS constraints. These constraints include avoiding high-risk areas, preventing collisions with obstacles and other drones, and maintaining reliable connectivity with ground control. By leveraging RL, the proposed algorithm aims to enable adaptive decision-making and autonomous navigation in dynamic and complex environments, thereby enhancing the scalability and efficiency of drone swarm operations in BVLOS scenarios.
Federated learning (FL) is a decentralized machine learning approach that enables multiple clients to collaboratively train a model without sharing their raw data. This approach is particularly beneficial for privacy-sensitive applications.
A thesis on FL systems with battery-constrained clients could explore novel algorithms and strategies for client selection, taking into account factors such as battery levels, data quality, and communication costs. The goal would be to optimize the trade-off between model performance, energy efficiency, and client participation, ensuring a robust and sustainable FL ecosystem.
The eternal vertex cover problem is a variant of the classical vertex cover problem defined in terms of an infinite attacker–defender game played on a graph. In each round of the game, the defender reconfigures guards from one vertex cover to another in response to a move by the attacker.
This thesis will encompass the development of theoretical frameworks and algorithmic approaches to tackle the EVC problem efficiently by examining the characteristics of specific graph classes.
Implementative Thesis: I also offer on the same subject a thesis focused only on the creation of a user-friendly EVC visualizer application. This application needs to provide a platform for users to explore the dynamics of the EVC problem through interactive visualization tools. Users will be able to input graphs of various sizes and structures and observe how different strategies unfold in real-time. The visualizer will support features such as step-by-step solution visualization, graph manipulation tools, and customizable parameters for experimentation.
This thesis focus on the task of online mining frequent itemsets, a crucial problem in data mining with applications in various domains such as e-commerce data analysis and web usage mining. The objective is to devise a novel algorithm capable of efficiently identifying frequent itemsets from streaming data while also implementing and benchmarking existing algorithms from the literature.