GLoT: A Novel Gated-Logarithmic Transformer for Efficient Sign Language Translation

Emirates Center for Mobility Research UAE University, Al-Ain, United Arab Emirates Correspondence: Leila@uaeu.ac.ae Abstract - Machine Translation has played a critical role in Language Machine Translation (SLMT) has been less explored. In this The number of Deaf and Hard of Hearing (DHH) paper, we aim to address this void by proposing the Gated population is expected to double to 860 million by 2050 [1]. Logarithmic Transformer (GLoT), which introduces a In addition to the existence of more than hundreds of sign gating mechanism that selectively filters out irrelevant languages [2], and an acute shortage of sign language information, ensuring that only the most critical temporal interpreters [3], there is a pressing need for automated and dependencies are retained [16]. By incorporating precise Sign Language Machine Translation (SLMT) logarithmic transformations, GLoT is designed to better systems. Having an inclusive communication could be capture long-range temporal patterns, improving the lifesaving in a tragic event such as a medical emergency.