Regularización con cotas de Lipschitz
En esta charla se introduce un metodo para reducir el sobreajuste en redes neuronales usando una regularización adaptativa basada en cotas de Lipschitz (LBA).
Professional Summary
I am currently a Ph.D. Candidate in Engineering at Universidad de los Andes (Colombia) focusing on multi-agent AI systems and cooperative resilience property. I began this program in 2023, supported by a competitive scholarship (awarded by Google DeepMind). I hold a Master’s degree in Applied Mathematics, which I completed in 2022. My thesis centered on neural networks and theoretical approaches to mitigating overfitting, combining mathematical principles with practical AI applications. Prior to that, I earned a Bachelor’s degree in Electronic Engineering in 2018, where I developed a strong foundation in systems design, programming, and signal processing.
Education
Ph.D. in Engineering
Universidad de los Andes, Colombia
M.Sc. in Applied Mathematics
Universidad Nacional de Colombia
B.Sc. in Electronic Engineering
Universidad de Nariño
Interests
Cooperative AI
Multi-Agent Systems
Inverse Reinforcement Learning
Resilience in Complex Systems
Neural Networks
Machine Learning
Autonomous Agents
📚 My Research
My research focuses on understanding and designing resilient multi-agent systems, particularly mixed-motive settings such as social dilemmas. I use inverse reinforcement learning (IRL)—with an emphasis on preference-based reward inference—to reveal latent social objectives and leverage the notion of cooperative resilience to guide reward learning that sustains collaboration under disruption (e.g., resource scarcity, failures, and strategic/adversarial behavior). The goal is to develop agents that adapt, maintain performance and fairness, and thrive despite shocks and non-stationarity—ultimately aligning individual incentives with collective well-being and long-term stability.
Please reach out to collaborate 😃
First-author Publications
(2025).
Evaluating Cooperative Resilience in Multiagent Systems: A Comparison Between Humans and LLMs.
arXiv.
(2025).
Reducing overfitting in ResNet with Adaptive Lipschitz regularization.
Journal of Computational and Applied Mathematics.
(2025).
Cooperative Resilience in Artificial Intelligence Multi-Agent Systems.
IEEE Transactions on Artificial Intelligence.
Co-authored publications
(2026).
Photocatalytic reduction of Cr(VI) in wastewater using TiO₂ coatings synthesized by plasma electrolytic oxidation (PEO): Influence of duty cycle and EDTA addition.
Next Materials.
(2025).
Evaluating Generalization Capabilities of LLM-Based Agents in Mixed-Motive Scenarios Using Concordia.
NeurIPS 2025 Datasets and Benchmarks Track.
Recent & Upcoming Talks
Resiliencia en sistemas multiagente y la experiencia de un doctorado
En esta charla hablaremos sobre mi investigación actual en sistemas multiagente de intereses mixtos y la experiencia en el camino del doctorado.
Blogs
🧠 NeurIPS 2025: Notes from a Shy Girl in Machine Learning
A personal reflection on attending NeurIPS 2025, highlighting the scientific environment and the human experience of a big AI conference.
📚 Recommended Readings on Cooperative AI (Vol. 1)
Cooperative AI is not only a technical challenge but also a question about how to design systems that balance individual incentives and collective welfare. In this first edition, I …
🤖 Reflections from the Cooperative AI Summer School 2025
Highlights from my experience at the Cooperative AI Foundation Summer School — where researchers from around the world gathered to discuss how cooperation and AI can help address …