Emanuele La Malfa

Emanuele La Malfa is a Researcher at the University of Oxford's Department of Computer Science, affiliated with Trinity College. His research focuses on robustness guarantees in Natural Language Processing (NLP) and model explainability, addressing critical challenges in AI fairness, adversarial robustness, and theoretical foundations. He is involved in the FUN2MODEL project and has contributed to understanding linguistic structure fragility in neural networks.

Key research themes include evaluating LLM robustness against adversarial attacks, exploring ethical implications of model exploitation (e.g., jailbreaking), and analyzing dialect fairness in reasoning tasks. His work bridges theoretical computer science with practical applications in AI safety and model interpretability.

Publications span topics like multi-agent system limitations in LLMs, fixed-point explainability frameworks, and formalizing continuity properties of language models. His research emphasizes scalable communication protocols for distributed AI systems and the logical capabilities of large models through representation learning.

La Malfa's contributions highlight interdisciplinary efforts, combining mathematical rigor with real-world NLP challenges, positioning him at the forefront of modern AI reliability research.