Public defence in Computer Science, M.Sc. (Tech) Suhas Thejaswi Muniyappa
Opponent: Dr. Senior Lecturer Othon Michail, University of Liverpool, England
Custos: Professor Aristides Gionis, Aalto University School of Science, Department of Computer Science
The public defence will be organised on campus and via Zoom. Link to the event
Thesis available for public display at: https://aaltodoc.aalto.fi/doc_public/eonly/riiputus/
Electronic thesis can be found at: https://aaltodoc.aalto.fi/handle/123456789/117017
Public defence announcement:
Many real-world problem instances are large, for example, a transportation network consists of thousands of transit points along with hundreds of thousands of transit options to choose. To further complicate, travelers have budget, travel-time and transit-time restrictions which makes a real-world problem increasingly complicated to solve. There exist a multitude of such problems that need to be solved in our day-to-day lives for which we rely on algorithmic decision making systems. In one hand, many algorithmic solutions which has a theoretical bound on the quality of the solution returned fail to perform in practice when the problem instances are huge, on the other hand, many heuristic solutions which can solve large problem instances fail to ensure a theoretical bound on the quality of solution returned. This thesis studies the intersection of both aspects of algorithm design principles, that is, to design scalable algorithmic solutions that also have a theoretical guarantee on the solution returned.
This thesis contributes to the design of scalable algorithmic frameworks that can be applied to solve important day-to-day problems such as contact-tracing in epidemic-contact networks, route planning for travel recommendations, avoiding bias in algorithmic decision making systems and quantifying the price paid to achieve algorithmic fairness, among others. We pursue our goal by developing an end-to-end research methodology which can be summarized under four main categories: first, we introduce novel problem formulations in the field of pattern detection and algorithmic fairness; second, we analyze the computational complexity of problems that we introduce, which help us differentiate between the problems that can and cannot be solved efficiently; third, we design and engineer algorithms with provable-theoretical guarantees while emphasizing the empirical scalability of algorithms to massive real-world datasets; and finally, we perform exhaustive experiments to demonstrate that the proposed solutions are practical.
Contact details of the doctoral student: [email protected], 0505230101