This is an internal, incomplete preview of a proposed change to the ACL Anthology.
For efficiency reasons, we generate only three BibTeX files per volume, and the preview may be incomplete in other ways, or contain mistakes.
Do not treat this content as an official publication.
Traditional lectures have poorer outcomes compared to active learning methodologies, yet many natural language processing classes in higher education still follow this outdated methodology. In this paper, we present, co-creational teaching, a methodology that encourages partnership between staff and lecturers and show how this can be applied to teach natural language processing. As a fast-moving and dynamic area of study with high interest from students, natural language processing is an ideal subject for innovative teaching methodologies to improve student outcomes. We detail our experience with teaching natural language processing through partnership with students and provide detailed descriptions of methodologies that can be used by others in their teaching, including considerations of diverse student populations.
The development of ontologies in various languages is attracting attention as the amount of multilingual data available on the web increases. Cross-lingual ontology matching facilitates interoperability amongst ontologies in different languages. Although supervised machine learning-based methods have shown good performance on ontology matching, their application to the cross-lingual setting is limited by the availability of training data. Current state-of-the-art unsupervised methods for cross-lingual ontology matching focus on lexical similarity between entities. These approaches follow a two-stage pipeline where the entities are translated into a common language using a translation service in the first step followed by computation of lexical similarity between the translations to match the entities in the second step. In this paper we introduce a novel ontology matching method based on the fusion of structural similarity and cross-lingual semantic similarity. We carry out experiments using 3 language pairs and report substantial improvements on the performance of the lexical methods thus showing the effectiveness of our proposed approach. To the best of our knowledge this is the first work which tackles the problem of unsupervised ontology matching in the cross-lingual setting by leveraging both structural and semantic embeddings.
This paper presents the development of CHAMUÇA, a novel lexical resource designed to document the influence of the Portuguese language on various Asian languages, with an initial focus on the languages of South Asia. Through the utilization of linked open data and the OntoLex vocabulary, CHAMUÇA offers structured insights into the linguistic characteristics, and cultural ramifications of Portuguese borrowings across multiple languages. The article outlines CHAMUÇA’s potential contributions to the linguistic linked data community, emphasising its role in addressing the scarcity of resources for lesser-resourced languages and serving as a test case for organising etymological data in a queryable format. CHAMUÇA emerges as an initiative towards the comprehensive catalogization and analysis of Portuguese borrowings, offering valuable insights into language contact dynamics, historical evolution, and cultural exchange in Asia, one that is based on linked data technology.
Corpus data is the main source of data for natural language processing applications, however no standard or model for corpus data has become predominant in the field. Linguistic linked data aims to provide methods by which data can be made findable, accessible, interoperable and reusable (FAIR). However, current attempts to create a linked data format for corpora have been unsuccessful due to the verbose and specialised formats that they use. In this work, we present the Teanga data model, which uses a layered annotation model to capture all NLP-relevant annotations. We present the YAML serializations of the model, which is concise and uses a widely-deployed format, and we describe how this can be interpreted as RDF. Finally, we demonstrate three examples of the use of the Teanga data model for syntactic annotation, literary analysis and multilingual corpora.
Text style transfer (TST) involves altering the linguistic style of a text while preserving its style-independent content. This paper focuses on sentiment transfer, a popular TST subtask, across a spectrum of Indian languages: Hindi, Magahi, Malayalam, Marathi, Punjabi, Odia, Telugu, and Urdu, expanding upon previous work on English-Bangla sentiment transfer. We introduce dedicated datasets of 1,000 positive and 1,000 negative style-parallel sentences for each of these eight languages. We then evaluate the performance of various benchmark models categorized into parallel, non-parallel, cross-lingual, and shared learning approaches, including the Llama2 and GPT-3.5 large language models (LLMs). Our experiments highlight the significance of parallel data in TST and demonstrate the effectiveness of the Masked Style Filling (MSF) approach in non-parallel techniques. Moreover, cross-lingual and joint multilingual learning methods show promise, offering insights into selecting optimal models tailored to the specific language and task requirements. To the best of our knowledge, this work represents the first comprehensive exploration of the TST task as sentiment transfer across a diverse set of languages.
This paper reports on the shared tasks organized by the 21st IWSLT Conference. The shared tasks address 7 scientific challenges in spoken language translation: simultaneous and offline translation, automatic subtitling and dubbing, speech-to-speech translation, dialect and low-resource speech translation, and Indic languages. The shared tasks attracted 17 teams whose submissions are documented in 27 system papers. The growing interest towards spoken language translation is also witnessed by the constantly increasing number of shared task organizers and contributors to the overview paper, almost evenly distributed across industry and academia.
This paper discusses the organisation and findings of the SIGTYP 2024 Shared Task on Word Embedding Evaluation for Ancient and Historical Languages. The shared task was split into the constrained and unconstrained tracks and involved solving either 3 or 5 problems for either 13 or 16 ancient and historical languages belonging to 4 language families, and making use of 6 different scripts. There were 14 registrations in total, of which 3 teams submitted to each track. Out of these 6 submissions, 2 systems were successful in the constrained setting and another 2 in the uncon- strained setting, and 4 system description papers were submitted by different teams. The best average result for morphological feature prediction was about 96%, while the best average results for POS-tagging and lemmatisation were 96% and 94% respectively. At the word level, the winning team could not achieve a higher average accuracy across all 16 languages than 5.95%, which demonstrates the difficulty of this problem. At the character level, the best average result over 16 languages 55.62%
This paper describes the structure and findings of the WILDRE 2024 shared task on Code-mixed Less-resourced Sentiment Analysis for Indo-Aryan Languages. The participants were asked to submit the test data’s final prediction on CodaLab. A total of fourteen teams registered for the shared task. Only four participants submitted the system for evaluation on CodaLab, with only two teams submitting the system description paper. While all systems show a rather promising performance, they outperform the baseline scores.
POS-tagging is typically considered a fundamental text preprocessing task, with a variety of downstream NLP tasks and techniques being dependent on the availability of POS-tagged corpora. As such, POS-taggers are important precursors to further NLP tasks, and their accuracy can impact the potential accuracy of these dependent tasks. While a variety of POS-tagging methods have been developed which work well with modern languages, historical languages present orthographic and editorial challenges which require special attention. The effectiveness of POS-taggers developed for modern languages is reduced when applied to Old Irish, with its comparatively complex orthography and morphology. This paper examines some of the obstacles to POS-tagging Old Irish text, and shows that inconsistencies between extant annotated corpora reduce the quantity of data available for use in training POS-taggers. The development of a multi-layer neural network model for POS-tagging Old Irish text is described, and an experiment is detailed which demonstrates that this model outperforms a variety of off-the-shelf POS-taggers. Moreover, this model sets a new benchmark for POS-tagging diplomatically edited Old Irish text.
The present paper introduces new sentiment data, MaCMS, for Magahi-Hindi-English (MHE) code-mixed language, where Magahi is a less-resourced minority language. This dataset is the first Magahi-Hindi-English code-mixed dataset for sentiment analysis tasks. Further, we also provide a linguistics analysis of the dataset to understand the structure of code-mixing and a statistical study to understand the language preferences of speakers with different polarities. With these analyses, we also train baseline models to evaluate the dataset’s quality.
Exploiting cognates for transfer learning in under-resourced languages is an exciting opportunity for language understanding tasks, including unsupervised machine translation, named entity recognition and information retrieval. Previous approaches mainly focused on supervised cognate detection tasks based on orthographic, phonetic or state-of-the-art contextual language models, which under-perform for most under-resourced languages. This paper proposes a novel language-agnostic weakly-supervised deep cognate detection framework for under-resourced languages using morphological knowledge from closely related languages. We train an encoder to gain morphological knowledge of a language and transfer the knowledge to perform unsupervised and weakly-supervised cognate detection tasks with and without the pivot language for the closely-related languages. While unsupervised, it overcomes the need for hand-crafted annotation of cognates. We performed experiments on different published cognate detection datasets across language families and observed not only significant improvement over the state-of-the-art but also our method outperformed the state-of-the-art supervised and unsupervised methods. Our model can be extended to a wide range of languages from any language family as it overcomes the requirement of the annotation of the cognate pairs for training.
We propose a new dataset for detecting non-inclusive language in sentences in English. These sentences were gathered from public sites, explaining what is inclusive and what is non-inclusive. We also extracted potentially non-inclusive keywords/phrases from the guidelines from business websites. A phrase dictionary was created by using an automatic extension with a word embedding trained on a massive corpus of general English text. In the end, a phrase dictionary was constructed by hand-editing the previous one to exclude inappropriate expansions and add the keywords from the guidelines. In a business context, the words individuals use can significantly impact the culture of inclusion and the quality of interactions with clients and prospects. Knowing the right words to avoid helps customers of different backgrounds and historically excluded groups feel included. They can make it easier to have productive, engaging, and positive communications. You can find the dictionaries, the code, and the method for making requests for the corpus at (we will release the link for data and code once the paper is accepted).
This article describes the manual construction of a part of the Old English WordNet (Old-EWN) covering the semantic field of emotion terms. This manually constructed part of the wordnet is to be eventually integrated with the automatically generated/manually checked part covering the whole of the rest of the Old English lexicon (currently under construction). We present the workflow for the definition of these emotion synsets on the basis of a dataset produced by a specialist in this area. We also look at the enrichment of the original Global WordNet Association Lexical Markup Framework (GWA LMF) schema to include the extra information which this part of the OldEWN requires. In the final part of the article we discuss how the wordnet style of lexicon organisation can be used to share and disseminate research findings/datasets in lexical semantics.
In this project note we describe our work to make better documentation for the Open Multilingual Wordnet (OMW), a platform integrating many open wordnets. This includes the documentation of the OMW website itself as well as of semantic relations used by the component wordnets. Some of this documentation work was done with the support of the Google Season of Docs. The OMW project page, which links both to the actual OMW server and the documentation has been moved to a new location: https://omwn.org.
This paper describes the structure and findings of the SIGTYP 2023 shared task on cognate and derivative detection for low-resourced languages, broken down into a supervised and unsupervised sub-task. The participants were asked to submit the test data’s final prediction. A total of nine teams registered for the shared task where seven teams registered for both sub-tasks. Only two participants ended up submitting system descriptions, with only one submitting systems for both sub-tasks. While all systems show a rather promising performance, all could be within the baseline score for the supervised sub-task. However, the system submitted for the unsupervised sub-task outperforms the baseline score.
Homophobia and Transphobia Detection is the task of identifying homophobia, transphobia, and non-anti-LGBT+ content from the given corpus. Homophobia and transphobia are both toxic languages directed at LGBTQ+ individuals that are described as hate speech. This paper summarizes our findings on the “Homophobia and Transphobia Detection in social media comments” shared task held at LT-EDI 2022 - ACL 2022 1. This shared taskfocused on three sub-tasks for Tamil, English, and Tamil-English (code-mixed) languages. It received 10 systems for Tamil, 13 systems for English, and 11 systems for Tamil-English. The best systems for Tamil, English, and Tamil-English scored 0.570, 0.870, and 0.610, respectively, on average macro F1-score.
Hope Speech detection is the task of classifying a sentence as hope speech or non-hope speech given a corpus of sentences. Hope speech is any message or content that is positive, encouraging, reassuring, inclusive and supportive that inspires and engenders optimism in the minds of people. In contrast to identifying and censoring negative speech patterns, hope speech detection is focussed on recognising and promoting positive speech patterns online. In this paper, we report an overview of the findings and results from the shared task on hope speech detection for Tamil, Malayalam, Kannada, English and Spanish languages conducted in the second workshop on Language Technology for Equality, Diversity and Inclusion (LT-EDI-2022) organised as a part of ACL 2022. The participants were provided with annotated training & development datasets and unlabelled test datasets in all the five languages. The goal of the shared task is to classify the given sentences into one of the two hope speech classes. The performances of the systems submitted by the participants were evaluated in terms of micro-F1 score and weighted-F1 score. The datasets for this challenge are openly available
Automatic Term Extraction (ATE) is one of the core problems in natural language processing and forms a key component of text mining pipelines of domain specific corpora. Complex low-level tasks such as machine translation and summarization for domain specific texts necessitate the use of term extraction systems. However, the development of these systems requires the use of large annotated datasets and thus there has been little progress made on this front for under-resourced languages. As a part of ongoing research, we present a dataset for term extraction from Hindi texts in this paper. To the best of our knowledge, this is the first dataset that provides term annotated documents for Hindi. Furthermore, we have evaluated this dataset on statistical term extraction methods and the results obtained indicate the problems associated with development of term extractors for under-resourced languages.
Knowledge-grounded dialogue systems utilise external knowledge such as knowledge graphs to generate informative and appropriate responses. A crucial challenge of such systems is to select facts from a knowledge graph pertinent to the dialogue context for response generation. This fact selection can be formulated as path traversal over a knowledge graph conditioned on the dialogue context. Such paths can originate from facts mentioned in the dialogue history and terminate at the facts to be mentioned in the response. These walks, in turn, provide an explanation of the flow of the conversation. This work proposes KG-CRuSE, a simple, yet effective LSTM based decoder that utilises the semantic information in the dialogue history and the knowledge graph elements to generate such paths for effective conversation explanation. Extensive evaluations showed that our model outperforms the state-of-the-art models on the OpenDialKG dataset on multiple metrics.
This paper introduces a new Magahi-Hindi-English (MHE) code-mixed data-set for similar language identification (SMLID), where Magahi is a less-resourced minority language. This corpus provides a language id at two levels: word and sentence. This data-set is the first Magahi-Hindi-English code-mixed data-set for similar language identification task. Furthermore, we will discuss the complexity of the data-set and provide a few baselines for the language identification task.
In this paper we will discuss our preliminary work towards the construction of a WordNet for Old English, taking our inspiration from other similar WN construction projects for ancient languages such as Ancient Greek, Latin and Sanskrit. The Old English WordNet (OldEWN) will build upon this innovative work in a number of different ways which we articulate in the article, most importantly by treateating figurative meaning as a ‘first-class citizen’ in the structuring of the semantic system. From a more practical perspective we will describe our plan to utilize a pre-existing lexicographic resource and the naisc system to automatically compile a provisional version of the WordNet which will then be checked and enriched by Old English experts.
Language resources are a key component of natural language processing and related research and applications. Users of language resources have different needs in terms of format, language, topics, etc. for the data they need to use. Linghub (McCrae and Cimiano, 2015) was first developed for this purpose, using the capabilities of linked data to represent metadata, and tackling the heterogeneous metadata issue. Linghub aimed at helping language resources and technology users to easily find and retrieve relevant data, and identify important information on access, topics, etc. This work describes a rejuvenation and modernisation of the 2015 platform into using a popular open source data management system, DSpace, as foundation. The new platform, Linghub2, contains updated and extended resources, more languages offered, and continues the work towards homogenisation of metadata through conversions, through linkage to standardisation strategies and community groups, such as the Open Digital Rights Language (ODRL) community group.
Pharmaceutical text classification is an important area of research for commercial and research institutions working in the pharmaceutical domain. Addressing this task is challenging due to the need of expert verified labelled data which can be expensive and time consuming to obtain. Towards this end, we leverage predictive coding methods for the task as they have been shown to generalise well for sentence classification. Specifically, we utilise GAN-BERT architecture to classify pharmaceutical texts. To capture the domain specificity, we propose to utilise the BioBERT model as our BERT model in the GAN-BERT framework. We conduct extensive evaluation to show the efficacy of our approach over baselines on multiple metrics.
This paper presents an outline of the shared task on translation of under-resourced Dravidian languages at DravidianLangTech-2022 workshop to be held jointly with ACL 2022. A description of the datasets used, approach taken for analysis of submissions and the results have been illustrated in this paper. Five sub-tasks organized as a part of the shared task include the following translation pairs: Kannada to Tamil, Kannada to Telugu, Kannada to Sanskrit, Kannada to Malayalam and Kannada to Tulu. Training, development and test datasets were provided to all participants and results were evaluated on the gold standard datasets. A total of 16 research groups participated in the shared task and a total of 12 submission runs were made for evaluation. Bilingual Evaluation Understudy (BLEU) score was used for evaluation of the translations.
This paper presents the development of the Parallel Universal Dependency (PUD) Treebank for two Indo-Aryan languages: Bengali and Magahi. A treebank of 1,000 sentences has been created using a parallel corpus of English and the UD framework. A preliminary set of sentences was annotated manually - 600 for Bengali and 200 for Magahi. The rest of the sentences were built using the Bengali and Magahi parser. The sentences have been translated and annotated manually by the authors, some of whom are also native speakers of the languages. The objective behind this work is to build a syntactically-annotated linguistic repository for the aforementioned languages, that can prove to be a useful resource for building further NLP tools. Additionally, Bengali and Magahi parsers were also created which is built on machine learning approach. The accuracy of the Bengali parser is 78.13% in the case of UPOS; 76.99% in the case of XPOS, 56.12% in the case of UAS; and 47.19% in the case of LAS. The accuracy of Magahi parser is 71.53% in the case of UPOS; 66.44% in the case of XPOS, 58.05% in the case of UAS; and 33.07% in the case of LAS. This paper also includes an illustration of the annotation schema followed, the findings of the Parallel Universal Dependency (PUD) treebank, and it’s resulting linguistic analysis
Social media platforms such as Twitter and Facebook have been utilised for various research studies, from the cohort-level discussion to community-driven approaches to address the challenges in utilizing social media data for health, clinical and biomedical information. Detection of medical jargon’s, named entity recognition, multi-word expression becomes the primary, fundamental steps in solving those challenges. In this paper, we enumerate the ULD-NUIG team’s system, designed as part of Social Media Mining for Health Applications (#SMM4H) Shared Task 2021. The team conducted a series of experiments to explore the challenges of task 6 and task 5. The submitted systems achieve F-1 0.84 and 0.53 score for task 6 and 5 respectively.
The application of predictive coding techniques to legal texts has the potential to greatly reduce the cost of legal review of documents, however, there is such a wide array of legal tasks and continuously evolving legislation that it is hard to construct sufficient training data to cover all cases. In this paper, we investigate few-shot and zero-shot approaches that require substantially less training data and introduce a triplet architecture, which for promissory statements produces performance close to that of a supervised system. This method allows predictive coding methods to be rapidly developed for new regulations and markets.
Words are defined based on their meanings in various ways in different resources. Aligning word senses across monolingual lexicographic resources increases domain coverage and enables integration and incorporation of data. In this paper, we explore the application of classification methods using manually-extracted features along with representation learning techniques in the task of word sense alignment and semantic relationship detection. We demonstrate that the performance of classification methods dramatically varies based on the type of semantic relationships due to the nature of the task but outperforms the previous experiments.
The Global Wordnet Formats have been introduced to enable wordnets to have a common representation that can be integrated through the Global WordNet Grid. As a result of their adoption, a number of shortcomings of the format were identified, and in this paper we describe the extensions to the formats that address these issues. These include: ordering of senses, dependencies between wordnets, pronunciation, syntactic modelling, relations, sense keys, metadata and RDF support. Furthermore, we provide some perspectives on how these changes help in the integration of wordnets.
WordNet is the most widely used lexical resource for English, while Wikidata is one of the largest knowledge graphs of entity and concepts available. While, there is a clear difference in the focus of these two resources, there is also a significant overlap and as such a complete linking of these resources would have many uses. We propose the development of such a linking, first by means of the hapax legomenon links and secondly by the use of natural language processing techniques. We show that these can be done with high accuracy but that human validation is still necessary. This has resulted in over 9,000 links being added between these two resources.
Multilingual sentence embeddings capture rich semantic information not only for measuring similarity between texts but also for catering to a broad range of downstream cross-lingual NLP tasks. State-of-the-art multilingual sentence embedding models require large parallel corpora to learn efficiently, which confines the scope of these models. In this paper, we propose a novel sentence embedding framework based on an unsupervised loss function for generating effective multilingual sentence embeddings, eliminating the need for parallel corpora. We capture semantic similarity and relatedness between sentences using a multi-task loss function for training a dual encoder model mapping different languages onto the same vector space. We demonstrate the efficacy of an unsupervised as well as a weakly supervised variant of our framework on STS, BUCC and Tatoeba benchmark tasks. The proposed unsupervised sentence embedding framework outperforms even supervised state-of-the-art methods for certain under-resourced languages on the Tatoeba dataset and on a monolingual benchmark. Further, we show enhanced zero-shot learning capabilities for more than 30 languages, with the model being trained on only 13 languages. Our model can be extended to a wide range of languages from any language family, as it overcomes the requirement of parallel corpora for training.
This paper presents an overview of the shared task on machine translation of Dravidian languages. We presented the shared task results at the EACL 2021 workshop on Speech and Language Technologies for Dravidian Languages. This paper describes the datasets used, the methodology used for the evaluation of participants, and the experiments’ overall results. As a part of this shared task, we organized four sub-tasks corresponding to machine translation of the following language pairs: English to Tamil, English to Malayalam, English to Telugu and Tamil to Telugu which are available at https://competitions.codalab.org/competitions/27650. We provided the participants with training and development datasets to perform experiments, and the results were evaluated on unseen test data. In total, 46 research groups participated in the shared task and 7 experimental runs were submitted for evaluation. We used BLEU scores for assessment of the translations.
Detecting offensive language in social media in local languages is critical for moderating user-generated content. Thus, the field of offensive language identification in under-resourced Tamil, Malayalam and Kannada languages are essential. As the user-generated content is more code-mixed and not well studied for under-resourced languages, it is imperative to create resources and conduct benchmarking studies to encourage research in under-resourced Dravidian languages. We created a shared task on offensive language detection in Dravidian languages. We summarize here the dataset for this challenge which are openly available at https://competitions.codalab.org/competitions/27654, and present an overview of the methods and the results of the competing systems.
Bilingual lexicons are a vital tool for under-resourced languages and recent state-of-the-art approaches to this leverage pretrained monolingual word embeddings using supervised or semi-supervised approaches. However, these approaches require cross-lingual information such as seed dictionaries to train the model and find a linear transformation between the word embedding spaces. Especially in the case of low-resourced languages, seed dictionaries are not readily available, and as such, these methods produce extremely weak results on these languages. In this work, we focus on the Dravidian languages, namely Tamil, Telugu, Kannada, and Malayalam, which are even more challenging as they are written in unique scripts. To take advantage of orthographic information and cognates in these languages, we bring the related languages into a single script. Previous approaches have used linguistically sub-optimal measures such as the Levenshtein edit distance to detect cognates, whereby we demonstrate that the longest common sub-sequence is linguistically more sound and improves the performance of bilingual lexicon induction. We show that our approach can increase the accuracy of bilingual lexicon induction methods on these languages many times, making bilingual lexicon induction approaches feasible for such under-resourced languages.
Automatic Language Identification (LI) or Dialect Identification (DI) of short texts of closely related languages or dialects, is one of the primary steps in many natural language processing pipelines. Language identification is considered a solved task in many cases; however, in the case of very closely related languages, or in an unsupervised scenario (where the languages are not known in advance), performance is still poor. In this paper, we propose the Unsupervised Deep Language and Dialect Identification (UDLDI) method, which can simultaneously learn sentence embeddings and cluster assignments from short texts. The UDLDI model understands the sentence constructions of languages by applying attention to character relations which helps to optimize the clustering of languages. We have performed our experiments on three short-text datasets for different language families, each consisting of closely related languages or dialects, with very minimal training sets. Our experimental evaluations on these datasets have shown significant improvement over state-of-the-art unsupervised methods and our model has outperformed state-of-the-art LI and DI systems in supervised settings.
Conversational recommender systems focus on the task of suggesting products to users based on the conversation flow. Recently, the use of external knowledge in the form of knowledge graphs has shown to improve the performance in recommendation and dialogue systems. Information from knowledge graphs aids in enriching those systems by providing additional information such as closely related products and textual descriptions of the items. However, knowledge graphs are incomplete since they do not contain all factual information present on the web. Furthermore, when working on a specific domain, knowledge graphs in its entirety contribute towards extraneous information and noise. In this work, we study several subgraph construction methods and compare their performance across the recommendation task. We incorporate pre-trained embeddings from the subgraphs along with positional embeddings in our models. Extensive experiments show that our method has a relative improvement of at least 5.62% compared to the state-of-the-art on multiple metrics on the recommendation task.
Social media are interactive platforms that facilitate the creation or sharing of information, ideas or other forms of expression among people. This exchange is not free from offensive, trolling or malicious contents targeting users or communities. One way of trolling is by making memes, which in most cases combines an image with a concept or catchphrase. The challenge of dealing with memes is that they are region-specific and their meaning is often obscured in humour or sarcasm. To facilitate the computational modelling of trolling in the memes for Indian languages, we created a meme dataset for Tamil (TamilMemes). We annotated and released the dataset containing suspected trolls and not-troll memes. In this paper, we use the a image classification to address the difficulties involved in the classification of troll memes with the existing methods. We found that the identification of a troll meme with such an image classifier is not feasible which has been corroborated with precision, recall and F1-score.
NUIG-Panlingua-KMI submission to WMT 2020 seeks to push the state-of-the-art in Similar Language Translation Task for Hindi↔Marathi language pair. As part of these efforts, we conducteda series of experiments to address the challenges for translation between similar languages. Among the 4 MT systems prepared under this task, 1 PBSMT systems were prepared for Hindi↔Marathi each and 1 NMT systems were developed for Hindi↔Marathi using Byte PairEn-coding (BPE) into subwords. The results show that different architectures NMT could be an effective method for developing MT systems for closely related languages. Our Hindi-Marathi NMT system was ranked 8th among the 14 teams that participated and our Marathi-Hindi NMT system was ranked 8th among the 11 teams participated for the task.
Code mixing is a common phenomena in multilingual societies where people switch from one language to another for various reasons. Recent advances in public communication over different social media sites have led to an increase in the frequency of code-mixed usage in written language. In this paper, we present the Generative Morphemes with Attention (GenMA) Model sentiment analysis system contributed to SemEval 2020 Task 9 SentiMix. The system aims to predict the sentiments of the given English-Hindi code-mixed tweets without using word-level language tags instead inferring this automatically using a morphological model. The system is based on a novel deep neural network (DNN) architecture, which has outperformed the baseline F1-score on the test data-set as well as the validation data-set. Our results can be found under the user name “koustava” on the “Sentimix Hindi English” page.
This paper reports on an ongoing task of monolingual word sense alignment in which a comparative study between the Portuguese Academy of Sciences Dictionary and the Dicionário Aberto is carried out in the context of the ELEXIS (European Lexicographic Infrastructure) project. Word sense alignment involves searching for matching senses within dictionary entries of different lexical resources and linking them, which poses significant challenges. The lexicographic criteria are not always entirely consistent within individual dictionaries and even less so across different projects where different options may have been assumed in terms of structure and especially wording techniques of lexicographic glosses. This hinders the task of matching senses. We aim to present our annotation workflow in Portuguese using the Semantic Web technologies. The results obtained are useful for the discussion within the community.
Hate speech detection in social media communication has become one of the primary concerns to avoid conflicts and curb undesired activities. In an environment where multilingual speakers switch among multiple languages, hate speech detection becomes a challenging task using methods that are designed for monolingual corpora. In our work, we attempt to analyze, detect and provide a comparative study of hate speech in a code-mixed social media text. We also provide a Hindi-English code-mixed data set consisting of Facebook and Twitter posts and comments. Our experiments show that deep learning models trained on this code-mixed corpus perform better.
In this paper we discuss the experience of bringing together over 40 different wordnets. We introduce some extensions to the GWA wordnet LMF format proposed in Vossen et al. (2016) and look at how this new information can be displayed. Notable extensions include: confidence, corpus frequency, orthographic variants, lexicalized and non-lexicalized synsets and lemmas, new parts of speech, and more. Many of these extensions already exist in multiple wordnets – the challenge was to find a compatible representation. To this end, we introduce a new version of the Open Multilingual Wordnet (Bond and Foster, 2013), that integrates a new set of tools that tests the extensions introduced by this new format, while also ensuring the integrity of the Collaborative Interlingual Index (CILI: Bond et al., 2016), avoiding the same new concept to be introduced through multiple projects.
Aligning senses across resources and languages is a challenging task with beneficial applications in the field of natural language processing and electronic lexicography. In this paper, we describe our efforts in manually aligning monolingual dictionaries. The alignment is carried out at sense-level for various resources in 15 languages. Moreover, senses are annotated with possible semantic relationships such as broadness, narrowness, relatedness, and equivalence. In comparison to previous datasets for this task, this dataset covers a wide range of languages and resources and focuses on the more challenging task of linking general-purpose language. We believe that our data will pave the way for further advances in alignment and evaluation of word senses by creating new solutions, particularly those notoriously requiring data such as neural networks. Our resources are publicly available at https://github.com/elexis-eu/MWSA.
In this paper we describe the contributions made by the European H2020 project “Prêt-à-LLOD” (‘Ready-to-use Multilingual Linked Language Data for Knowledge Services across Sectors’) to the further development of the Linguistic Linked Open Data (LLOD) infrastructure. Prêt-à-LLOD aims to develop a new methodology for building data value chains applicable to a wide range of sectors and applications and based around language resources and language technologies that can be integrated by means of semantic technologies. We describe the methods implemented for increasing the number of language data sets in the LLOD. We also present the approach for ensuring interoperability and for porting LLOD data sets and services to other infrastructures, as well as the contribution of the projects to existing standards.
Metaphor comprehension and understanding is a complex cognitive task that requires interpreting metaphors by grasping the interaction between the meaning of their target and source concepts. This is very challenging for humans, let alone computers. Thus, automatic metaphor interpretation is understudied in part due to the lack of publicly available datasets. The creation and manual annotation of such datasets is a demanding task which requires huge cognitive effort and time. Moreover, there will always be a question of accuracy and consistency of the annotated data due to the subjective nature of the problem. This work addresses these issues by presenting an annotation scheme to interpret verb-noun metaphoric expressions in text. The proposed approach is designed with the goal of reducing the workload on annotators and maintain consistency. Our methodology employs an automatic retrieval approach which utilises external lexical resources, word embeddings and semantic similarity to generate possible interpretations of identified metaphors in order to enable quick and accurate annotation. We validate our proposed approach by annotating around 1,500 metaphors in tweets which were annotated by six native English speakers. As a result of this work, we publish as linked data the first gold standard dataset for metaphor interpretation which will facilitate research in this area.
This paper presents a bidirectional transformer based approach for recognising semantic relationships between a pair of words as proposed by CogALex VI shared task in 2020. The system presented here works by employing BERT embeddings of the words and passing the same over tuned neural network to produce a learning model for the pair of words and their relationships. Afterwards the very same model is used for the relationship between unknown words from the test set. CogALex VI provided Subtask 1 as the identification of relationship of three specific categories amongst English pair of words and the presented system opts to work on that. The resulted relationships of the unknown words are analysed here which shows a balanced performance in overall characteristics with some scope for improvement.
There is an increasing demand for sentiment analysis of text from social media which are mostly code-mixed. Systems trained on monolingual data fail for code-mixed data due to the complexity of mixing at different levels of the text. However, very few resources are available for code-mixed data to create models specific for this data. Although much research in multilingual and cross-lingual sentiment analysis has used semi-supervised or unsupervised methods, supervised methods still performs better. Only a few datasets for popular languages such as English-Spanish, English-Hindi, and English-Chinese are available. There are no resources available for Malayalam-English code-mixed data. This paper presents a new gold standard corpus for sentiment analysis of code-mixed text in Malayalam-English annotated by voluntary annotators. This gold standard corpus obtained a Krippendorff’s alpha above 0.8 for the dataset. We use this new corpus to provide the benchmark for sentiment analysis in Malayalam-English code-mixed texts.
Understanding the sentiment of a comment from a video or an image is an essential task in many applications. Sentiment analysis of a text can be useful for various decision-making processes. One such application is to analyse the popular sentiments of videos on social media based on viewer comments. However, comments from social media do not follow strict rules of grammar, and they contain mixing of more than one language, often written in non-native scripts. Non-availability of annotated code-mixed data for a low-resourced language like Tamil also adds difficulty to this problem. To overcome this, we created a gold standard Tamil-English code-switched, sentiment-annotated corpus containing 15,744 comment posts from YouTube. In this paper, we describe the process of creating the corpus and assigning polarities. We present inter-annotator agreement and show the results of sentiment analysis trained on this corpus as a benchmark.
The OntoLex vocabulary enjoys increasing popularity as a means of publishing lexical resources with RDF and as Linked Data. The recent publication of a new OntoLex module for lexicography, lexicog, reflects its increasing importance for digital lexicography. However, not all aspects of digital lexicography have been covered to the same extent. In particular, supplementary information drawn from corpora such as frequency information, links to attestations, and collocation data were considered to be beyond the scope of lexicog. Therefore, the OntoLex community has put forward the proposal for a novel module for frequency, attestation and corpus information (FrAC), that not only covers the requirements of digital lexicography, but also accommodates essential data structures for lexical information in natural language processing. This paper introduces the current state of the OntoLex-FrAC vocabulary, describes its structure, some selected use cases, elementary concepts and fundamental definitions, with a focus on frequency and attestations.
In this paper, we present the NUIG system at the TIAD shard task. This system includes graph-based metrics calculated using novel algorithms, with an unsupervised document embedding tool called ONETA and an unsupervised multi-way neural machine translation method. The results are an improvement over our previous system and produce the highest precision among all systems in the task as well as very competitive F-Measure results. Incorporating features from other systems should be easy in the framework we describe in this paper, suggesting this could very easily be extended to an even stronger result.
Identifying metaphors in text is very challenging and requires comprehending the underlying comparison. The automation of this cognitive process has gained wide attention lately. However, the majority of existing approaches concentrate on word-level identification by treating the task as either single-word classification or sequential labelling without explicitly modelling the interaction between the metaphor components. On the other hand, while existing relation-level approaches implicitly model this interaction, they ignore the context where the metaphor occurs. In this work, we address these limitations by introducing a novel architecture for identifying relation-level metaphoric expressions of certain grammatical relations based on contextual modulation. In a methodology inspired by works in visual reasoning, our approach is based on conditioning the neural network computation on the deep contextualised features of the candidate expressions using feature-wise linear modulation. We demonstrate that the proposed architecture achieves state-of-the-art results on benchmark datasets. The proposed methodology is generic and could be applied to other textual classification problems that benefit from contextual interaction.
WordNet, while one of the most widely used resources for NLP, has not been updated for a long time, and as such a new project English WordNet has arisen to continue the development of the model under an open-source paradigm. In this paper, we detail the second release of this resource entitled “English WordNet 2020”. The work has focused firstly, on the introduction of new synsets and senses and developing guidelines for this and secondly, on the integration of contributions from other projects. We present the changes in this edition, which total over 15,000 changes over the previous release.
Metaphor processing and understanding has attracted the attention of many researchers recently with an increasing number of computational approaches. A common factor among these approaches is utilising existing benchmark datasets for evaluation and comparisons. The availability, quality and size of the annotated data are among the main difficulties facing the growing research area of metaphor processing. The majority of current approaches pertaining to metaphor processing concentrate on word-level processing due to data availability. On the other hand, approaches that process metaphors on the relation-level ignore the context where the metaphoric expression. This is due to the nature and format of the available data. Word-level annotation is poorly grounded theoretically and is harder to use in downstream tasks such as metaphor interpretation. The conversion from word-level to relation-level annotation is non-trivial. In this work, we attempt to fill this research gap by adapting three benchmark datasets, namely the VU Amsterdam metaphor corpus, the TroFi dataset and the TSV dataset, to suit relation-level metaphor identification. We publish the adapted datasets to facilitate future research in relation-level metaphor processing.
In this paper we describe the current state of development of the Linguistic Linked Open Data (LLOD) infrastructure, an LOD(sub-)cloud of linguistic resources, which covers various linguistic data bases, lexicons, corpora, terminology and metadata repositories. We give in some details an overview of the contributions made by the European H2020 projects “Prêt-à-LLOD” (‘Ready-to-useMultilingual Linked Language Data for Knowledge Services across Sectors’) and “ELEXIS” (‘European Lexicographic Infrastructure’) to the further development of the LLOD.
With regard to the wider area of AI/LT platform interoperability, we concentrate on two core aspects: (1) cross-platform search and discovery of resources and services; (2) composition of cross-platform service workflows. We devise five different levels (of increasing complexity) of platform interoperability that we suggest to implement in a wider federation of AI/LT platforms. We illustrate the approach using the five emerging AI/LT platforms AI4EU, ELG, Lynx, QURATOR and SPEAKER.
We describe the release of a new wordnet for English based on the Princeton WordNet, but now developed under an open-source model. In particular, this version of WordNet, which we call English WordNet 2019, which has been developed by multiple people around the world through GitHub, fixes many errors in previous wordnets for English. We give some details of the changes that have been made in this version and give some perspectives about likely future changes that will be made as this project continues to evolve.
Neologism detection is a key task in the constructing of lexical resources and has wider implications for NLP, however the identification of multiword neologisms has received little attention. In this paper, we show that we can effectively identify the distinction between compositional and non-compositional adjective-noun pairs by using pretrained language models and comparing this with individual word embeddings. Our results show that the use of these models significantly improves over baseline linguistic features, however the combination with linguistic features still further improves the results, suggesting the strength of a hybrid approach.
Lexical resource differ from encyclopaedic resources and represent two distinct types of resource covering general language and named entities respectively. However, many lexical resources, including Princeton WordNet, contain many proper nouns, referring to named entities in the world yet it is not possible or desirable for a lexical resource to cover all named entities that may reasonably occur in a text. In this paper, we propose that instead of including synsets for instance concepts PWN should instead provide links to Wikipedia articles describing the concept. In order to enable this we have created a gold-quality mapping between all of the 7,742 instances in PWN and Wikipedia (where such a mapping is possible). As such, this resource aims to provide a gold standard for link discovery, while also allowing PWN to distinguish itself from other resources such as DBpedia or BabelNet. Moreover, this linking connects PWN to the Linguistic Linked Open Data cloud, thus creating a richer, more usable resource for natural language processing.
Wordnets are extensively used in natural language processing, but the current approaches for manually building a wordnet from scratch involves large research groups for a long period of time, which are typically not available for under-resourced languages. Even if wordnet-like resources are available for under-resourced languages, they are often not easily accessible, which can alter the results of applications using these resources. Our proposed method presents an expand approach for improving and generating wordnets with the help of machine translation. We apply our methods to improve and extend wordnets for the Dravidian languages, i.e., Tamil, Telugu, Kannada, which are severly under-resourced languages. We report evaluation results of the generated wordnet senses in term of precision for these languages. In addition to that, we carried out a manual evaluation of the translations for the Tamil language, where we demonstrate that our approach can aid in improving wordnet resources for under-resourced Dravidian languages.
The paper describes objectives, concept and methodology for ELEXIS, a European infrastructure fostering cooperation and information exchange among lexicographical research communities. The infrastructure is a newly granted project under the Horizon 2020 INFRAIA call, with the topic Integrating Activities for Starting Communities. The project is planned to start in January 2018.
Princeton WordNet is one of the most widely-used resources for natural language processing, but is updated only infrequently and cannot keep up with the fast-changing usage of the English language on social media platforms such as Twitter. The Colloquial WordNet aims to provide an open platform whereby anyone can contribute, while still following the structure of WordNet. Many crowd-sourced lexical resources often have significant quality issues, and as such care must be taken in the design of the interface to ensure quality. In this paper, we present the development of a platform that can be opened on the Web to any lexicographer who wishes to contribute to this resource and the lexicographic methodology applied by this interface.
Metaphor is an essential element of human cognition which is often used to express ideas and emotions that might be difficult to express using literal language. Processing metaphoric language is a challenging task for a wide range of applications ranging from text simplification to psychotherapy. Despite the variety of approaches that are trying to process metaphor, there is still a need for better models that mimic the human cognition while exploiting fewer resources. In this paper, we present an approach based on distributional semantics to identify metaphors on the phrase-level. We investigated the use of different word embeddings models to identify verb-noun pairs where the verb is used metaphorically. Several experiments are conducted to show the performance of the proposed approach on benchmark datasets.
This paper describes the construction and annotation of a corpus of verbal MWEs for English, as part of the PARSEME Shared Task 1.1 on automatic identification of verbal MWEs. The criteria for corpus selection, the categories of MWEs used, and the training process are discussed, along with the particular issues that led to revisions in edition 1.1 of the annotation guidelines. Finally, an overview of the characteristics of the final annotated corpus is presented, as well as some discussion on inter-annotator agreement.
The Open Linguistics Working Group (OWLG) brings together researchers from various fields of linguistics, natural language processing, and information technology to present and discuss principles, case studies, and best practices for representing, publishing and linking linguistic data collections. A major outcome of our work is the Linguistic Linked Open Data (LLOD) cloud, an LOD (sub-)cloud of linguistic resources, which covers various linguistic databases, lexicons, corpora, terminologies, and metadata repositories. We present and summarize five years of progress on the development of the cloud and of advancements in open data in linguistics, and we describe recent community activities. The paper aims to serve as a guideline to orient and involve researchers with the community and/or Linguistic Linked Open Data.
Princeton WordNet is one of the most important resources for natural language processing, but is only available for English. While it has been translated using the expand approach to many other languages, this is an expensive manual process. Therefore it would be beneficial to have a high-quality automatic translation approach that would support NLP techniques, which rely on WordNet in new languages. The translation of wordnets is fundamentally complex because of the need to translate all senses of a word including low frequency senses, which is very challenging for current machine translation approaches. For this reason we leverage existing translations of WordNet in other languages to identify contextual information for wordnet senses from a large set of generic parallel corpora. We evaluate our approach using 10 translated wordnets for European languages. Our experiment shows a significant improvement over translation without any contextual information. Furthermore, we evaluate how the choice of pivot languages affects performance of multilingual word sense disambiguation.
This paper introduces the motivation for and design of the Collaborative InterLingual Index (CILI). It is designed to make possible coordination between multiple loosely coupled wordnet projects. The structure of the CILI is based on the Interlingual index first proposed in the EuroWordNet project with several pragmatic extensions: an explicit open license, definitions in English and links to wordnets in the Global Wordnet Grid.
In this paper, we describe a new and improved Global Wordnet Grid that takes advantage of the Collaborative InterLingual Index (CILI). Currently, the Open Multilingal Wordnet has made many wordnets accessible as a single linked wordnet, but as it used the Princeton Wordnet of English (PWN) as a pivot, it loses concepts that are not part of PWN. The technical solution to this, a central registry of concepts, as proposed in the EuroWordnet project through the InterLingual Index, has been known for many years. However, the practical issues of how to host this index and who decides what goes in remained unsolved. Inspired by current practice in the Semantic Web and the Linked Open Data community, we propose a way to solve this issue. In this paper we define the principles and protocols for contributing to the Grid. We tested them on two use cases, adding version 3.1 of the Princeton WordNet to a CILI based on 3.0 and adding the Open Dutch Wordnet, to validate the current set up. This paper aims to be a call for action that we hope will be further discussed and ultimately taken up by the whole wordnet community.
Recent years have witnessed a surge in the amount of semantic information published on the Web. Indeed, the Web of Data, a subset of the Semantic Web, has been increasing steadily in both volume and variety, transforming the Web into a ‘global database’ in which resources are linked across sites. Linguistic fields – in a broad sense – have not been left behind, and we observe a similar trend with the growth of linguistic data collections on the so-called ‘Linguistic Linked Open Data (LLOD) cloud’. While both Semantic Web and Natural Language Processing communities can obviously take advantage of this growing and distributed linguistic knowledge base, they are today faced with a new challenge, i.e., that of facilitating multilingual access to the Web of data. In this paper we present the publication of BabelNet 2.0, a wide-coverage multilingual encyclopedic dictionary and ontology, as Linked Data. The conversion made use of lemon, a lexicon model for ontologies particularly well-suited for this enterprise. The result is an interlinked multilingual (lexical) resource which can not only be accessed on the LOD, but also be used to enrich existing datasets with linguistic information, or to support the process of mapping datasets across languages.
The creation of language resources is a time-consuming process requiring the efforts of many people. The use of resources collaboratively created by non-linguistists can potentially ameliorate this situation. However, such resources often contain more errors compared to resources created by experts. For the particular case of lexica, we analyse the case of Wiktionary, a resource created along wiki principles and argue that through the use of a principled lexicon model, namely Lemon, the resulting data could be better understandable to machines. We then present a platform called Lemon Source that supports the creation of linked lexical data along the Lemon model. This tool builds on the concept of a semantic wiki to enable collaborative editing of the resources by many users concurrently. In this paper, we describe the model, the tool and present an evaluation of its usability based on a small group of users.