This study focuses on identifying highly inclined muons in water-Cherenkov detectors similar to those used by the Pierre Auger Observatory using neural networks. Highly inclined muons, which are distinctive signatures of air showers induced by neutrinos or cosmic rays arriving at significant inclinations, offer a lower background rate compared to less inclined atmospheric particles. We explore the transition from conventional statistical approaches to machine learning methodologies to identify highly inclined muons by leveraging their unique signatures in the temporal signal distributions of three photosensors that uniformly observe the volume of a water-Cherenkov detector. By adopting machine learning, particularly neural network techniques, we aim to enhance the identification of highly inclined muons, thus improving triggering schemas designed for detecting neutrino primaries. This study not only advances the identification of highly inclined muons but also investigates the optimization of machine learning models for their efficient recognition within the water-Cherenkov detector setup.