Louis Annabi

Centre National de la Recherche Scientifique, École Nationale Supérieure de Techniques Avancées

Papers

Total Citations

H-Index

About

Louis Annabi is a researcher at the forefront of embodied AI and continual learning, with a focus on enabling robots to learn and adapt from human demonstrations in real time. Their work bridges the gap between neuroscience-inspired algorithms and practical robotics, particularly through the application of predictive coding for sequential modeling. In their 2022 paper on "Continual Sequence Modeling With Predictive Coding" (5 citations), Annabi proposed an alternative to backpropagation-through-time, allowing recurrent neural networks to learn continuously without batch training—a critical step toward lifelong learning in dynamic environments. More recently, their 2024 study on "Unsupervised Motion Retargeting for Human-Robot Imitation" (4 citations) tackles the fundamental challenge of translating human motion sequences into robot-compatible actions, accounting for differences in embodiment. This early-stage work leverages deep learning to enable online, unsupervised imitation, paving the way for more intuitive human-robot collaboration. Though still early in their career, Annabi’s contributions are already shaping how machines can learn from and mirror human behavior, making their research essential reading for anyone interested in the intersection of continual learning, motion generation, and interactive robotics.