Marwan Elghitany

Papers

Total Citations

H-Index

About

Marwan Elghitany is a researcher advancing the frontier of automatic speech recognition (ASR), with a particular focus on low-resource languages. His most-cited work, "Advancing Arabic Speech Recognition Through Large-Scale Weakly Supervised Learning" (2025, 3 citations), tackles the critical challenge of building high-performing ASR models for languages like Arabic, which suffer from limited annotated data. By leveraging large-scale weakly supervised learning, Elghitany’s approach reduces reliance on expensive, manually transcribed datasets, enabling more robust and scalable speech-to-text systems for applications ranging from conversational agents to automated subtitling. This contribution is especially impactful for Arabic, a morphologically rich and dialectally diverse language, where traditional supervised methods fall short. Though early in his career, Elghitany’s work signals a promising trajectory in democratizing ASR technology for underserved languages. His research not only addresses a pressing technical gap but also holds practical implications for industrial robotics, call center automation, and human-machine interaction in Arabic-speaking regions. As the demand for inclusive AI grows, Elghitany’s innovations in weakly supervised learning position him as a key contributor to making speech technology more accessible and equitable worldwide.