Text simplification resources for Spanish

Proceedings of The 9th Linguistic Annotation Workshop

“…Yet, the "passivization" rule may still be productive in other textual typologies, where it can happen that the author of the simplification prefers not only to keep, but even to insert, a passive, in order to avoid more unusual syntactic constructs in Italian (such as impersonal sentences). This is also in line with what Bott and Saggion (2014) …”

“…Recently, ATS has been used in educational scenarios and assistive technologies; e.g. for the adaptation of texts to particular readers, like children (De Belder et al, 2010), L2 learners (Petersen andOstendorf, 2007), people with low literacy skills (Aluísio et al, 2008), cognitive disabilities (Bott and Saggion, 2014) or language impairments, e.g. aphasia (Carroll et al, 1998) or deafness (Inui et al, 2003).…”

“…For some of these languages, parallel monolingual corpora are annotated with simplification rules corresponding to transformations to perform on a complex sentence. This is the approach followed by Brouwers et al (2014) for French; Bott and Saggion (2014) (2010) for Brazilian Portuguese, who adopted phrase-based machine learning from a parallel corpus. For Basque, Aranzabe et al (2013) used the output of a readability assessment system for detecting complex sentences, which are simplified by a large set of hand-crafted rules.…”

The corpus of Basque simplified texts (CBST)

“…This is done by annotating the changes performed when simplifying. To our knowledge, the operations performed in the simplification are only presented in the case of the Brazilian Portuguese corpus (Caseli et al 2009), the Spanish corpus (Bott and Saggion 2014) and the Italian corpus (Brunato et al 2015) but only in the cases of the Spanish and Italian corpora, these operations are organised in annotation schemes.…”

“…The cue words of the operation we are describing in each case will be underlined in both cases (Basque and English, as mentioned before). The different operations presented in this scheme are based on the annotation of the corpus; the structure of the annotation scheme has also been compared to the Spanish (Bott and Saggion 2014) and the Italian (Brunato et al 2015) annotation schemes.…”