Bilal FAYE

🏆 Best Academic Paper Award – EGC 2024

January 2024 – EGC Conference

Awarded for the paper “Context Normalization”, presented at the EGC 2024 conference. This work introduces a novel normalization approach leveraging prior knowledge to build context-based structures. Unlike Batch Normalization, which depends on mini-batch statistics and assumes naive data distributions, Context Normalization operates within predefined contexts, offering more stability and better performance in real-world scenarios.

Proposed as a modular layer, Context Normalization can be seamlessly integrated into various points in neural architectures, facilitating more adaptive and robust model training.

Learn More