Workshop on Linguistic Resources for NLP (2018)


up

pdf (full)
bib (full)
Proceedings of the First Workshop on Linguistic Resources for Natural Language Processing

pdf bib
Proceedings of the First Workshop on Linguistic Resources for Natural Language Processing
Peter Machonis | Anabela Barreiro | Kristina Kocijan | Max Silberztein

pdf bib
Corpus Phonetics: Past, Present, and Future
Mark Liberman

Invited talk

pdf bib
Using Linguistic Resources to Evaluate the Quality of Annotated Corpora
Max Silberztein

Statistical and neural-network-based methods that compute their results by comparing a given text to be analyzed with a reference corpus assume that the reference corpus is complete and reliable enough. In this article, I conduct several experiments on an extract of the Open American National Corpus to verify this assumption.

pdf bib
Rule-based vs. Neural Net Approaches to Semantic Textual Similarity
Linrui Zhang | Dan Moldovan

This paper presents a neural net approach to determine Semantic Textual Similarity (STS) using attention-based bidirectional Long Short-Term Memory Networks (Bi-LSTM). To this date, most of the traditional STS systems were rule-based that built on top of excessive use of linguistic features and resources. In this paper, we present an end-to-end attention-based Bi-LSTM neural network system that solely takes word-level features, without expensive feature engineering work or the usage of external resources. By comparing its performance with traditional rule-based systems against SemEval-2012 benchmark, we make an assessment on the limitations and strengths of neural net systems to rule-based systems on Semantic Textual Similarity.

pdf bib
Linguistic Resources for Phrasal Verb Identification
Peter Machonis

This paper shows how a Lexicon-Grammar dictionary of English phrasal verbs (PV) can be transformed into an electronic dictionary, and with the help of multiple grammars, dictionaries, and filters within the linguistic development environment, NooJ, how to accurately identify PV in large corpora. The NooJ program is an alternative to statistical methods commonly used in NLP: all PV are listed in a dictionary and then located by means of a PV grammar in both continuous and discontinuous format. Results are then refined with a series of dictionaries, disambiguating grammars, and other linguistics recourses. The main advantage of such a program is that all PV can be identified in any corpus. The only drawback is that PV not listed in the dictionary (e.g., archaic forms, recent neologisms) are not identified; however, new PV can easily be added to the electronic dictionary, which is freely available to all.

pdf bib
Designing a Croatian Aspectual Derivatives Dictionary: Preliminary Stages
Kristina Kocijan | Krešimir Šojat | Dario Poljak

The paper focusses on derivationally connected verbs in Croatian, i.e. on verbs that share the same lexical morpheme and are derived from other verbs via prefixation, suffixation and/or stem alternations. As in other Slavic languages with rich derivational morphology, each verb is marked for aspect, either perfective or imperfective. Some verbs, mostly of foreign origin, are marked as bi-aspectual verbs. The main objective of this paper is to detect and to describe major derivational processes and affixes used in the derivation of aspectually connected verbs with NooJ. Annotated chains are exported into a format adequate for web database system and further used to enhance the aspectual and derivational information for each verb.

pdf bib
A Rule-Based System for Disambiguating French Locative Verbs and Their Translation into Arabic
Safa Boudhina | Héla Fehri

This paper presents a rule-based system for disambiguating frensh locative verbs and their translation to Arabic language. The disambiguation phase is based on the use of the French Verb dictionary (LVF) of Dubois and Dubois Charlier as a linguistic resource, from which a base of disambiguation rules is extracted. The extracted rules thus take the form of transducers which will be subsequently applied to texts. The translation phase consists in translating the disambiguated locative verbs returned by the disambiguation phase. The translation takes into account the verb’s tense used as well as the inflected form of the verb. This phase is based on bilingual dictionaries that contain the different French locative verbs and their translation into the Arabic language. The experimentation and the evaluation are done in the linguistic platform NooJ. The obtained results are satisfactory.

pdf bib
A Pedagogical Application of NooJ in Language Teaching: The Adjective in Spanish and Italian
Andrea Rodrigo | Mario Monteleone | Silvia Reyes

In this paper, a pedagogical application of NooJ to the teaching and learning of Spanish as a foreign language is presented, which is directed to a specific addressee: learners whose mother tongue is Italian. The category ‘adjective’ has been chosen on account of its lower frequency of occurrence in texts written in Spanish, and particularly in the Argentine Rioplatense variety, and with the aim of developing strategies to increase its use. In addition, the features that the adjective shares with other grammatical categories render it extremely productive and provide elements that enrich the learners’ proficiency. The reference corpus contains the front pages of the Argentinian newspaper Clarín related to an emblematic historical moment, whose starting point is 24 March 1976, when a military coup began, and covers a thirty year period until 24 March 2006. It can be seen how the term desaparecido emerges with all its cultural and social charge, providing a context which allows an approach to Rioplatense Spanish from a more comprehensive perspective. Finally, a pedagogical proposal accounting for the application of the NooJ platform in language teaching is included.

pdf bib
STYLUS: A Resource for Systematically Derived Language Usage
Bonnie Dorr | Clare Voss

We describe a resource derived through extraction of a set of argument realizations from an existing lexical-conceptual structure (LCS) Verb Database of 500 verb classes (containing a total of 9525 verb entries) to include information about realization of arguments for a range of different verb classes. We demonstrate that our extended resource, called STYLUS (SysTematicallY Derived Language USe), enables systematic derivation of regular patterns of language usage without requiring manual annotation. We posit that both spatially oriented applications such as robot navigation and more general applications such as narrative generation require a layered representation scheme where a set of primitives (often grounded in space/motion such as GO) is coupled with a representation of constraints at the syntax-semantics interface. We demonstrate that the resulting resource covers three cases of lexico-semantic operations applicable to both language understanding and language generation.

pdf bib
Contemporary Amharic Corpus: Automatically Morpho-Syntactically Tagged Amharic Corpus
Andargachew Mekonnen Gezmu | Binyam Ephrem Seyoum | Michael Gasser | Andreas Nürnberger

We introduced the contemporary Amharic corpus, which is automatically tagged for morpho-syntactic information. Texts are collected from 25,199 documents from different domains and about 24 million orthographic words are tokenized. Since it is partly a web corpus, we made some automatic spelling error correction. We have also modified the existing morphological analyzer, HornMorpho, to use it for the automatic tagging.

pdf bib
Gold Corpus for Telegraphic Summarization
Chanakya Malireddy | Srivenkata N M Somisetty | Manish Shrivastava

Most extractive summarization techniques operate by ranking all the source sentences and then select the top ranked sentences as the summary. Such methods are known to produce good summaries, especially when applied to news articles and scientific texts. However, they don’t fare so well when applied to texts such as fictional narratives, which don’t have a single central or recurrent theme. This is because usually the information or plot of the story is spread across several sentences. In this paper, we discuss a different summarization technique called Telegraphic Summarization. Here, we don’t select whole sentences, rather pick short segments of text spread across sentences, as the summary. We have tailored a set of guidelines to create such summaries and, using the same, annotate a gold corpus of 200 English short stories.

pdf bib
Design of a Tigrinya Language Speech Corpus for Speech Recognition
Hafte Abera | Sebsibe H/Mariam

In this paper, we describe the first Tigrinya Languages speech corpora designed and development for speech recognition purposes. Tigrinya, often written as Tigrigna (ትግርኛ) /tɪˈɡrinjə/ belongs to the Semitic branch of the Afro-Asiatic languages where it shows the characteristic features of a Semitic language. It is spoken by ethnic Tigray-Tigrigna people in the Horn of Africa. The paper outlines different corpus designing process analysis of related work on speech corpora creation for different languages. The authors provide also procedures that were used for the creation of Tigrinya speech recognition corpus which is the under-resourced language. One hundred and thirty speakers, native to Tigrinya language, were recorded for training and test dataset set. Each speaker read 100 texts, which consisted of syllabically rich and balanced sentences. Ten thousand sets of sentences were used to prompt sheets. These sentences contained all of the contextual syllables and phones.

pdf bib
Parallel Corpora for bi-Directional Statistical Machine Translation for Seven Ethiopian Language Pairs
Solomon Teferra Abate | Michael Melese | Martha Yifiru Tachbelie | Million Meshesha | Solomon Atinafu | Wondwossen Mulugeta | Yaregal Assabie | Hafte Abera | Binyam Ephrem | Tewodros Abebe | Wondimagegnhue Tsegaye | Amanuel Lemma | Tsegaye Andargie | Seifedin Shifaw

In this paper, we describe the development of parallel corpora for Ethiopian Languages: Amharic, Tigrigna, Afan-Oromo, Wolaytta and Geez. To check the usability of all the corpora we conducted baseline bi-directional statistical machine translation (SMT) experiments for seven language pairs. The performance of the bi-directional SMT systems shows that all the corpora can be used for further investigations. We have also shown that the morphological complexity of the Ethio-Semitic languages has a negative impact on the performance of the SMT especially when they are target languages. Based on the results we obtained, we are currently working towards handling the morphological complexities to improve the performance of statistical machine translation among the Ethiopian languages.

pdf bib
Using Embeddings to Compare FrameNet Frames Across Languages
Jennifer Sikos | Sebastian Padó

Much interest in Frame Semantics is fueled by the substantial extent of its applicability across languages. At the same time, lexicographic studies have found that the applicability of individual frames can be diminished by cross-lingual divergences regarding polysemy, syntactic valency, and lexicalization. Due to the large effort involved in manual investigations, there are so far no broad-coverage resources with “problematic” frames for any language pair. Our study investigates to what extent multilingual vector representations of frames learned from manually annotated corpora can address this need by serving as a wide coverage source for such divergences. We present a case study for the language pair English — German using the FrameNet and SALSA corpora and find that inferences can be made about cross-lingual frame applicability using a vector space model.

pdf bib
Construction of a Multilingual Corpus Annotated with Translation Relations
Yuming Zhai | Aurélien Max | Anne Vilnat

Translation relations, which distinguish literal translation from other translation techniques, constitute an important subject of study for human translators (Chuquet and Paillard, 1989). However, automatic processing techniques based on interlingual relations, such as machine translation or paraphrase generation exploiting translational equivalence, have not exploited these relations explicitly until now. In this work, we present a categorisation of translation relations and annotate them in a parallel multilingual (English, French, Chinese) corpus of oral presentations, the TED Talks. Our long term objective will be to automatically detect these relations in order to integrate them as important characteristics for the search of monolingual segments in relation of equivalence (paraphrases) or of entailment. The annotated corpus resulting from our work will be made available to the community.

pdf bib
Towards an Automatic Classification of Illustrative Examples in a Large Japanese-French Dictionary Obtained by OCR
Christian Boitet | Mathieu Mangeot | Mutsuko Tomokiyo

We work on improving the Cesselin, a large and open source Japanese-French bilingual dictionary digitalized by OCR, available on the web, and contributively improvable online. Labelling its examples (about 226000) would significantly enhance their usefulness for language learners. Examples are proverbs, idiomatic constructions, normal usage examples, and, for nouns, phrases containing a quantifier. Proverbs are easy to spot, but not examples of other types. To find a method for automatically or at least semi-automatically annotating them, we have studied many entries, and hypothesized that the degree of lexical similarity between results of MT into a third language might give good cues. To confirm that hypothesis, we sampled 500 examples and used Google Translate to translate into English their Japanese expressions and their French translations. The hypothesis holds well, in particular for distinguishing examples of normal usage from idiomatic examples. Finally, we propose a detailed annotation procedure and discuss its future automatization.

pdf bib
Contractions: To Align or Not to Align, That Is the Question
Anabela Barreiro | Fernando Batista

This paper performs a detailed analysis on the alignment of Portuguese contractions, based on a previously aligned bilingual corpus. The alignment task was performed manually in a subset of the English-Portuguese CLUE4Translation Alignment Collection. The initial parallel corpus was pre-processed and, a decision was made as to whether the contraction should be maintained or decomposed in the alignment. Decomposition was required in the cases in which the two words that have been concatenated, i.e., the preposition and the determiner or pronoun, go in two separate translation alignment pairs (e.g., [no seio de] [a União Europeia] | [within] [the European Union]). Most contractions required decomposition in contexts where they are positioned at the end of a multiword unit. On the other hand, contractions tend to be maintained when they occur in the beginning or in the middle of the multiword unit, i.e., in the frozen part of the multiword (e.g., [no que diz respeito a] | [with regard to] or [além disso] [in addition]. A correct alignment of multiwords and phrasal units containing contractions is instrumental for machine translation, paraphrasing, and variety adaptation.

pdf bib
Enabling Code-Mixed Translation: Parallel Corpus Creation and MT Augmentation Approach
Mrinal Dhar | Vaibhav Kumar | Manish Shrivastava

Code-mixing, use of two or more languages in a single sentence, is ubiquitous; generated by multi-lingual speakers across the world. The phenomenon presents itself prominently in social media discourse. Consequently, there is a growing need for translating code-mixed hybrid language into standard languages. However, due to the lack of gold parallel data, existing machine translation systems fail to properly translate code-mixed text. In an effort to initiate the task of machine translation of code-mixed content, we present a newly created parallel corpus of code-mixed English-Hindi and English. We selected previously available English-Hindi code-mixed data as a starting point for the creation of our parallel corpus. We then chose 4 human translators, fluent in both English and Hindi, for translating the 6088 code-mixed English-Hindi sentences to English. With the help of the created parallel corpus, we analyzed the structure of English-Hindi code-mixed data and present a technique to augment run-of-the-mill machine translation (MT) approaches that can help achieve superior translations without the need for specially designed translation systems. We present an augmentation pipeline for existing MT approaches, like Phrase Based MT (Moses) and Neural MT, to improve the translation of code-mixed text. The augmentation pipeline is presented as a pre-processing step and can be plugged with any existing MT system, which we demonstrate by improving translations done by systems like Moses, Google Neural Machine Translation System (NMTS) and Bing Translator for English-Hindi code-mixed content.