Machine translation is the automated process of converting text or speech from one language to another using computer software, without a human translator directly performing the conversion. Early machine translation systems relied on rule-based grammar mappings and, later, on statistical models trained on large volumes of paired translated text.
Modern machine translation is dominated by neural machine translation, which uses transformer-based neural networks, the same core architecture behind large language models, to produce translations that capture context and nuance far better than earlier rule-based or purely statistical systems. Tools like Google Translate and the translation capabilities built into models like ChatGPT rely on this neural approach, which is why translation quality has improved so substantially over the past decade compared to older machine translation methods.
