Luisa Fernanda Estrada Plata

Contact

Luisa-Fernanda.Estrada-Plata@warwick.ac.uk

Office: D1.04 (Zeeman)

Side Quests

Outside my PhD life, I also train for a variety of endurance events: swimming, running, triathlon, obstacle races… You name it!

I am currently fundraising for Kidney Care UK as part of the Great North Swim, 5K challenge.

More info here.

About Me

I am a PhD student for the Mathematics for Real World Systems CDT working under the supervision of Paolo Turrini. I am interested in the use of Artificial Intelligence for social good. My research focuses on multi-agent systems and opinion dynamics on social networks using techniques from Game Theory, Graph Theory and Combinatorics.

I am also a seminar organiser for the Warwick SIAM-IMA chapter. We run the SPAAM seminar series and organise social events, hackathons, and the AMP conference. If you are interested in giving a talk, contact us at siam@warwick.ac.uk.

Publications

Estrada, L., Högele, M. and Steinicke, A., 2026. On the tradeoff between almost sure error tolerance and mean deviation frequency in martingale convergence.Link opens in a new window Latin American Journal of Probability and Mathematical Statistics. vol. XXII, p. 105-141. DOI 10.30757/ALEA.v23-06.
Chistikov, D., Estrada, L., Paterson, M. and Turrini, P., 2024. Learning a Social Network by Influencing Opinions. In Proceedings of the 23rd International Conference on Autonomous Agents and Multiagent Systems, AAMAS ’24, 363–371. Richland, SC: International Foundation for Autonomous Agents and Multiagent Systems. ISBN 9798400704864.
Estrada, L. and Högele, M., 2022. Moment estimates in the first Borel-Cantelli Lemma with applications to mean deviation frequencies. Statistics and Probability Letters. vol. 190. ISSN 0167-7152.

Pre-prints

Estrada, L., Glendinning, S. and Nugent, A., 2026. Speaking of Opinions: Comparing Approaches to Modelling Opinion Manipulation. Invited contribution to La Matematica: Official Journal of the Association for Women in Mathematics. (In review)

Estrada, L., Chistikov, D., Paterson, M. and Turrini, P., 2026. On the Limits of PAC Learning of Networks from Opinion Dynamics. Journal of Autonomous Agents and Multi-Agent Systems (In review)

Recorded Talks

I gave several talks at the SPAAM Seminar during my PhD; recordings are available here. If you would like to get in touch, feel free to contact me.

Learning a Social Network by Influencing Opinions

Overview: We study a campaigner who seeks to learn the structure of a social network by observing the underlying diffusion process and intervening on it. We focus on the dynamics of synchronous majority updates on binary opinions. We derive upper bounds on the budget required to learn any network with certainty, accounting for both observation and intervention resources, and further refine these bounds for clique networks. We also examine the campaigner's learning performance when the available budget falls below these bounds. In such cases, we propose a greedy intervention strategy that maximises information gain at each opinion diffusion step.

PAC Learning Social Networks

Overview: Agents in social networks change opinions when influenced by sufficiently many peers. Existing literature typically assumes that the network structure and dynamics are fully known, which is often unrealistic. We ask how hard it is to learn a network structure from samples of the agents' synchronous opinion updates. Not exactly, but in a Probably Approximately Correct (PAC) sense. The answer (spoiler): It is HARD. However, we propose a polynomial-time greedy algorithm that performs well in experiments across various classes of random graphs.