Caio Lima
  • Home
  • About
  • Projects

Caio Lima

Lab GitHub Lattes

Researcher in Probability, Statistical and Artificial Intelligence

Researcher Profile

I study probability theory, statistical inference and statistical learning theory, with emphasis in calculations theoretical, as a foundation for understanding modern machine learning systems.

My work is centered on the mathematical and statistical structure of artificial intelligence models. I am particularly interested in how learning systems can be derived, analyzed, and interpreted through rigorous deterministic or non-deterministic reasoning — including their assumptions, limitations, and expressive capabilities.

While my primary orientation is theoretical, I also develop and implement models in practice, connecting foundational principles with computational methods in machine learning and data analysis.

Research Areas

  • Probability Theory
  • Statistical Inference (Frequentist and Bayesian)
  • Statistical Learning
  • Artificial Neural Networks
  • Natural Language Processing
  • Generative Modeling (Generative AI)
  • Large Language Models (LLMs)
  • Multi-agent Systems

Research Questions

  • How can artificial neural networks be formally derived using probabilistic and statistical principles?

  • How can models based on deep neural networks be interpreted?

  • How can generative models be understood and interpreted in terms of latent variable structures and inference?

  • Can the estimated parameters of a LLM be interpreted for tasks applied in reality?

  • Can we measure the event known as “Hallucination” in an LLM?

  • Is current probability theory sufficient to provide an axiomatic basis for generative models?

B.Sc. in Statistics (in progress) | Federal University of Pará (UFPA)

Python R C SQL

Caio Lima ©