 What is the history of AI and translation? It goes back to the Cold War era where translators were made to translate from Russian to English. They were rule-based systems. From the 1980s, we started applying statistical models to language translation. This was done by solving two sub-problems. Translation modeling, which is to translate phrases, and then language modeling to make sure the phrases in the target language make sense as a translation. The 2000s saw the breakthrough of neural machine translation. These make use of neural networks to directly learn the task of translation, and they are just more accurate. Today, we make use of models like GPT, where a pre-trained language model is then fine-tuned on a translation task. So we need less translation-based data. Subscribe for more videos on artificial intelligence.