A Machine Translation Approach for Modernizing Historical Documents Using Backtranslation

A Machine Translation Approach for Modernizing Historical Documents Using Backtranslation Miguel Domingo, Francisco Casacuberta PRHLT Research Center Universitat Polit`ecnica de Val`encia midobal@prhlt.upv.es, fcn@prhlt.upv.es Abstract Human language evolves with the passage of time. This makes historical documents to be hard to comprehend by contemporary people and, thus, limits their accessi- bility to scholars specialized in the time period in which a certain document was written. Modernization aims at breaking this language barrier and increase the accessi- bility of historical documents to a broader audience. To do so, it generates a new version of a historical docu- ment, written in the modern version of the document’s original language. In this work, we propose several machine translation approaches for modernizing histori- cal documents. We tested these approaches in different scenarios, obtaining very encouraging results. 1. Introduction Historical documents are an important part of our cul- tural heritage. With the aim of their preservation, there is an increase need of digitalazing—creating a digital text version which can be searched and automatically processed— of historical documents [1]. However, there is an additional difficulty created by their linguistic prop- erties: On the one hand, human language evolves with the passage of time. On the other hand, due to a lack of a spelling convention, orthography changes depending on the author and time period in which a given document was written. These problems make historical documents harder to read, and makes it harder to digitalize them (since their digital text version needs to be searched and automatically processed). The orthography problem has been well researched in the literature [2, 3, 4, 5, 6, 7]. The proposed solu- tion that aims to solve this problem is known as spelling normalization, and its goal is to adapt the document’s spelling to modern standards in order to achieve an or- thography consistency and increase the document’s read- ability. However, while is true that spelling normaliza- tion makes historical documents easier to read, they are still hard to comprehend by contemporary people. This problem limits the accessibility of historical documents to scholars specialized in the time period in which the document was written. Modernization aims at breaking the language barrier, generating a new version of a historical document in the modern version of the language in which the docu- ment was originally written (see Fig. 1 for an example). Therefore, not only the orthography is updated. The lexicon and grammar are also modified in order to match the modern use of the document’s language. The main drawback of this solution is that part of the document’s original intention could be lost in the process (e.g., part of the rhyme in Fig. 1 is lost for the sake of clarity). Nonetheless, the document’s clarity is increased and, thus, its accessibility to a broader audience. To the best of our knowledge, modernization of historical documents has been less researched in the literature. A shared task was organized in order to translate historical text to contemporary language [9]. The shared task’s main goal was …

Content truncated to 500 words