Introducing rule-based machine translation
We will begin our journey of MT with the classical approach, known as rule-based machine translation (RBMT), which aims to exploit linguistic information about the source and target languages. RBMT techniques fall under the broad category of knowledge-based systems, which mainly aim to capture the knowledge of human experts to solve complex problems. For example, try to recall your first efforts in learning a foreign language. First, we had to find the correct translation of a sentence, which involved searching for it in a dictionary and mapping each word of the source sentence to a word in the target. Then, we had to make a few adjustments, such as finding the correct verb conjugation. Figure 6.3 illustrates this approach with an English sentence translated into French:
Figure 6.3 – A word-for-word mapping from the source (EN) to the target (FR) language
We can follow a similar approach and create word-for...