Max Robotics — Global Service Robot Directory + US Certification

About

Wendyam Eric Lionel Ilboudo is a researcher at the forefront of robust machine learning and robotics, specializing in developing optimization algorithms resilient to noisy data. His major contributions center on integrating Student-t distribution-based methods into deep learning frameworks, significantly advancing the field of robust optimization. His most cited work, "t-soft update of target network for deep reinforcement learning" (2021, 68 citations), introduces a novel approach to stabilize training in reinforcement learning environments. Ilboudo's foundational paper, "Robust Stochastic Gradient Descent With Student-t Distribution Based First-Order Momentum" (2020, 67 citations), demonstrates how to effectively handle outliers in training data, a critical challenge in real-world applications. He further refined these concepts with "TAdam: A Robust Stochastic Gradient Optimizer" (2020, 8 citations), a new optimizer that explicitly detects and discards outliers. His work on "Adaptive t-Momentum-based Optimization for Unknown Ratio of Outliers in Amateur Data in Imitation Learning" (2021, 3 citations) directly addresses the practical problem of transferring imperfect human demonstrations to robots. Ilboudo's research has garnered over 146 citations, establishing him as a key innovator in creating robust, noise-tolerant machine learning systems for robotics and beyond.