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.
Image-to-text generation involves automatically generating descriptive text from images and has applications in medical report generation. However, traditional approaches often exhibit a semantic gap between visual and textual information. In this paper, we propose a multi-task learning framework to leverage both visual and non-imaging data for generating radiology reports. Along with chest X-ray images, 10 additional features comprising numeric, binary, categorical, and text data were incorporated to create a unified representation. The model was trained to generate text, predict the degree of patient severity, and identify medical findings. Multi-task learning, especially with text generation prioritisation, improved performance over single-task baselines across language generation metrics. The framework also mitigated overfitting in auxiliary tasks compared to single-task models. Qualitative analysis showed logically coherent narratives and accurate identification of findings, though some repetition and disjointed phrasing remained. This work demonstrates the benefits of multi-modal, multi-task learning for image-to-text generation applications.
In spite of recent successes in improving Machine Translation (MT) quality overall, MT engines require a large amount of resources, which leads to markedly lower quality for lesser-resourced languages. This study explores the case of translation from English into Igbo, a very low resource language spoken by about 45 million speakers. With the aim of improving MT quality in this scenario, we investigate methods for guided detection of critical/harmful MT errors, more specifically those caused by non-compositional multi-word expressions and polysemy. We have designed diagnostic tests for these cases and applied them to collections of medical texts from CDC, Cochrane, NCDC, NHS and WHO.
While performance of many text classification tasks has been recently improved due to Pretrained Language Models (PLMs), in this paper we show that they still suffer from a performance gap when the underlying distribution of topics changes. For example, a genre classifier trained on political topics often fails when tested on documents in the same genre, but about sport or medicine. In this work, we quantify this phenomenon empirically with a large corpus and a large set of topics. Thus, we verify that domain transfer remains challenging both for classic PLMs, such as BERT, and for modern large models (LLMs), such as GPT. We develop a data augmentation approach by generating texts in any desired genre and on any desired topic, even when there are no documents in the training corpus that are both in that particular genre and on that particular topic. When we augment the training dataset with the topically-controlled synthetic texts, F1 improves up to 50% for some topics, approaching on-topic training, while showing no or next to no improvement for other topics. While our empirical results focus on genre classification, our methodology is applicable to other classification tasks such as gender, authorship, or sentiment classification.
We report our participation in the SemEval-2023 shared task on propaganda detection and describe our solutions with pre-trained models and their ensembles. For Subtask 1 (News Genre Categorisation), we report the impact of several settings, such as the choice of the classification models (monolingual or multilingual or their ensembles), the choice of the training sets (base or additional sources), the impact of detection certainty in making a classification decision as well as the impact of other hyper-parameters. In particular, we fine-tune models on additional data for other genre classification tasks, such as FTD. We also try adding texts from genre-homogenous corpora, such as Panorama, Babylon Bee for satire and Giganews for for reporting texts. We also make prepared models for Subtasks 2 and 3 with finetuning the corresponding models first for Subtask 1.The code needed to reproduce the experiments is available.
The paper presents the outcomes of AI-COVID19, our project aimed at better understanding of misinformation flow about COVID-19 across social media platforms. The specific focus of the study reported in this paper is on collecting data from Telegram groups which are active in promotion of COVID-related misinformation. Our corpus collected so far contains around 28 million words, from almost one million messages. Given that a substantial portion of misinformation flow in social media is spread via multimodal means, such as images and video, we have also developed a mechanism for utilising such channels via producing automatic transcripts for videos and automatic classification for images into such categories as memes, screenshots of posts and other kinds of images. The accuracy of the image classification pipeline is around 87%.
Genre identification is a kind of non-topic text classification. The main difference between this task and topic classification is that genre, unlike topic, usually cannot be expressed just by some keywords and is defined as a functional space. Neural models based on pre-trained transformers, such as BERT or XLM-RoBERTa, demonstrate SOTA results in many NLP tasks, including non-topical classification. However, in many cases, their downstream application to very large corpora, such as those extracted from social media, can lead to unreliable results because of dataset shifts, when some raw texts do not match the profile of the training set. To mitigate this problem, we experiment with individual models as well as with their ensembles. To evaluate the robustness of all models we use a prediction confidence metric, which estimates the reliability of a prediction in the absence of a gold standard label. We can evaluate robustness via the confidence gap between the correctly classified texts and the misclassified ones on a labeled test corpus, higher gaps make it easier to identify whether a text is classified correctly. Our results show that for all of the classifiers tested in this study, there is a confidence gap, but for the ensembles, the gap is wider, meaning that ensembles are more robust than their individual models.
Pre-trained transformer-based models, such as BERT, have shown excellent performance in most natural language processing benchmark tests, but we still lack a good understanding of the linguistic knowledge of BERT in Neural Machine Translation (NMT). Our work uses syntactic probes and Quality Estimation (QE) models to analyze the performance of BERT’s syntactic dependencies and their impact on machine translation quality, exploring what kind of syntactic dependencies are difficult for NMT engines based on BERT. While our probing experiments confirm that pre-trained BERT “knows” about syntactic dependencies, its ability to recognize them often decreases after fine-tuning for NMT tasks. We also detect a relationship between syntactic dependencies in three languages and the quality of their translations, which shows which specific syntactic dependencies are likely to be a significant cause of low-quality translations.
Current practices in building new NLP models for low-resourced languages rely either on Machine Translation of training sets from better resourced languages or on cross-lingual transfer from them. Still we can see a considerable performance gap between the models originally trained within better resourced languages and the models transferred from them. In this study we test the possibility of (1) using natural annotation to build synthetic training sets from resources not initially designed for the target downstream task and (2) employing curriculum learning methods to select the most suitable examples from synthetic training sets. We test this hypothesis across seven Slavic languages and across three curriculum learning strategies on Named Entity Recognition as the downstream task. We also test the possibility of fine-tuning the synthetic resources to reflect linguistic properties, such as the grammatical case and gender, both of which are important for the Slavic languages. We demonstrate the possibility to achieve the mean F1 score of 0.78 across the three basic entities types for Belarusian starting from zero resources in comparison to the baseline of 0.63 using the zero-shot transfer from English. For comparison, the English model trained on the original set achieves the mean F1-score of 0.75. The experimental results are available from https://github.com/ValeraLobov/SlavNER
This paper presents an attempt to build a Modern Standard Arabic (MSA) sentence-level simplification system. We experimented with sentence simplification using two approaches: (i) a classification approach leading to lexical simplification pipelines which use Arabic-BERT, a pre-trained contextualised model, as well as a model of fastText word embeddings; and (ii) a generative approach, a Seq2Seq technique by applying a multilingual Text-to-Text Transfer Transformer mT5. We developed our training corpus by aligning the original and simplified sentences from the internationally acclaimed Arabic novel Saaq al-Bambuu. We evaluate effectiveness of these methods by comparing the generated simple sentences to the target simple sentences using the BERTScore evaluation metric. The simple sentences produced by the mT5 model achieve P 0.72, R 0.68 and F-1 0.70 via BERTScore, while, combining Arabic-BERT and fastText achieves P 0.97, R 0.97 and F-1 0.97. In addition, we report a manual error analysis for these experiments.
In this paper, we present a Modern Standard Arabic (MSA) Sentence difficulty classifier, which predicts the difficulty of sentences for language learners using either the CEFR proficiency levels or the binary classification as simple or complex. We compare the use of sentence embeddings of different kinds (fastText, mBERT , XLM-R and Arabic-BERT), as well as traditional language features such as POS tags, dependency trees, readability scores and frequency lists for language learners. Our best results have been achieved using fined-tuned Arabic-BERT. The accuracy of our 3-way CEFR classification is F-1 of 0.80 and 0.75 for Arabic-Bert and XLM-R classification respectively and 0.71 Spearman correlation for regression. Our binary difficulty classifier reaches F-1 0.94 and F-1 0.98 for sentence-pair semantic similarity classifier.
The shared task of the 13th Workshop on Building and Using Comparable Corpora was devoted to the induction of bilingual dictionaries from comparable rather than parallel corpora. In this task, for a number of language pairs involving Chinese, English, French, German, Russian and Spanish, the participants were supposed to determine automatically the target language translations of several thousand source language test words of three frequency ranges. We describe here some background, the task definition, the training and test data sets and the evaluation used for ranking the participating systems. We also summarize the approaches used and present the results of the evaluation. In conclusion, the outcome of the competition are the results of a number of systems which provide surprisingly good solutions to the ambitious problem.
This paper explores the use of Deep Learning methods for automatic estimation of quality of human translations. Automatic estimation can provide useful feedback for translation teaching, examination and quality control. Conventional methods for solving this task rely on manually engineered features and external knowledge. This paper presents an end-to-end neural model without feature engineering, incorporating a cross attention mechanism to detect which parts in sentence pairs are most relevant for assessing quality. Another contribution concerns oprediction of fine-grained scores for measuring different aspects of translation quality, such as terminological accuracy or idiomatic writing. Empirical results on a large human annotated dataset show that the neural model outperforms feature-based methods significantly. The dataset and the tools are available.
This paper proposes a novel framework for digital curation of Web corpora in order to provide robust estimation of their parameters, such as their composition and the lexicon. In recent years language models pre-trained on large corpora emerged as clear winners in numerous NLP tasks, but no proper analysis of the corpora which led to their success has been conducted. The paper presents a procedure for robust frequency estimation, which helps in establishing the core lexicon for a given corpus, as well as a procedure for estimating the corpus composition via unsupervised topic models and via supervised genre classification of Web pages. The results of the digital curation study applied to several Web-derived corpora demonstrate their considerable differences. First, this concerns different frequency bursts which impact the core lexicon obtained from each corpus. Second, this concerns the kinds of texts they contain. For example, OpenWebText contains considerably more topical news and political argumentation in comparison to ukWac or Wikipedia. The tools and the results of analysis have been released.
Automatically recognizing an existing semantic relation (e.g. “is a”, “part of”, “property of”, “opposite of” etc.) between two words (phrases, concepts, etc.) is an important task affecting many NLP applications and has been subject of extensive experimentation and modeling. Current approaches to automatically telling if a relation exists between two given concepts X and Y can be grouped into two types: 1) those modeling word-paths connecting X and Y in text and 2) those modeling distributional properties of X and Y separately, not necessary in the proximity to each other. Here, we investigate how both types can be improved and combined. We suggest a distributional approach that is based on an attention-based transformer. We have also developed a novel word path model that combines useful properties of a convolutional network with a fully connected language model. While our transformer-based approach works better, both our models significantly outperform the state-of-the-art within their classes of approaches. We also demonstrate that combining the two approaches results in additional gains since they use somewhat different data sources.
The paper describes a computational approach to produce functionally comparable monolingual corpus resources for translation studies and contrastive analysis. We exploit a text-external approach, based on a set of Functional Text Dimensions to model text functions, so that each text can be represented as a vector in a multidimensional space of text functions. These vectors can be used to find reasonably homogeneous subsets of functionally similar texts across different corpora. Our models for predicting text functions are based on recurrent neural networks and traditional feature-based machine learning approaches. In addition to using the categories of the British National Corpus as our test case, we investigated the functional comparability of the English parts from the two parallel corpora: CroCo (English-German) and RusLTC (English-Russian) and applied our models to define functionally similar clusters in them. Our results show that the Functional Text Dimensions provide a useful description for text categories, while allowing a more flexible representation for texts with hybrid functions.
There is great variation in the amount of NLP resources available for Slavonic languages. For example, the Universal Dependency treebank (Nivre et al., 2016) has about 2 MW of training resources for Czech, more than 1 MW for Russian, while only 950 words for Ukrainian and nothing for Belorussian, Bosnian or Macedonian. Similarly, the Autodesk Machine Translation dataset only covers three Slavonic languages (Czech, Polish and Russian). In this talk I will discuss a general approach, which can be called Language Adaptation, similarly to Domain Adaptation. In this approach, a model for a particular language processing task is built by lexical transfer of cognate words and by learning a new feature representation for a lesser-resourced (recipient) language starting from a better-resourced (donor) language. More specifically, I will demonstrate how language adaptation works in such training scenarios as Translation Quality Estimation, Part-of-Speech tagging and Named Entity Recognition.
This paper presents the BUCC 2017 shared task on parallel sentence extraction from comparable corpora. It recalls the design of the datasets, presents their final construction and statistics and the methods used to evaluate system results. 13 runs were submitted to the shared task by 4 teams, covering three of the four proposed language pairs: French-English (7 runs), German-English (3 runs), and Chinese-English (3 runs). The best F-scores as measured against the gold standard were 0.84 (German-English), 0.80 (French-English), and 0.43 (Chinese-English). Because of the design of the dataset, in which not all gold parallel sentence pairs are known, these are only minimum values. We examined manually a small sample of the false negative sentence pairs for the most precise French-English runs and estimated the number of parallel sentence pairs not yet in the provided gold standard. Adding them to the gold standard leads to revised estimates for the French-English F-scores of at most +1.5pt. This suggests that the BUCC 2017 datasets provide a reasonable approximate evaluation of the parallel sentence spotting task.
In this paper we introduce MoBiL, a hybrid Monolingual, Bilingual and Language modelling feature set and feature selection and evaluation framework. The set includes translation quality indicators that can be utilized to automatically predict the quality of human translations in terms of content adequacy and language fluency. We compare MoBiL with the QuEst baseline set by using them in classifiers trained with support vector machine and relevance vector machine learning algorithms on the same data set. We also report an experiment on feature selection to opt for fewer but more informative features from MoBiL. Our experiments show that classifiers trained on our feature set perform consistently better in predicting both adequacy and fluency than the classifiers trained on the baseline feature set. MoBiL also performs well when used with both support vector machine and relevance vector machine algorithms.
Research in Natural Language Processing often relies on a large collection of manually annotated documents. However, currently there is no reliable genre-annotated corpus of web pages to be employed in Automatic Genre Identification (AGI). In AGI, documents are classified based on their genres rather than their topics or subjects. The major shortcoming of available web genre collections is their relatively low inter-coder agreement. Reliability of annotated data is an essential factor for reliability of the research result. In this paper, we present the first web genre corpus which is reliably annotated. We developed precise and consistent annotation guidelines which consist of well-defined and well-recognized categories. For annotating the corpus, we used crowd-sourcing which is a novel approach in genre annotation. We computed the overall as well as the individual categories’ chance-corrected inter-annotator agreement. The results show that the corpus has been annotated reliably.
The extraction of dictionaries from parallel text corpora is an established technique. However, as parallel corpora are a scarce resource, in recent years the extraction of dictionaries using comparable corpora has obtained increasing attention. In order to find a mapping between languages, almost all approaches suggested in the literature rely on a seed lexicon. The work described here achieves competitive results without requiring such a seed lexicon. Instead it presupposes mappings between comparable documents in different languages. For some common types of textual resources (e.g. encyclopedias or newspaper texts) such mappings are either readily available or can be established relatively easily. The current work is based on Wikipedias where the mappings between languages are determined by the authors of the articles. We describe a neural-network inspired algorithm which first characterizes each Wikipedia article by a number of keywords, and then considers the identification of word translations as a variant of word alignment in a noisy environment. We present results and evaluations for eight language pairs involving Germanic, Romanic, and Slavic languages as well as Chinese.
We present experiments in automatic genre classification on web corpora, comparing a wide variety of features on several different genreannotated datasets (HGC, I-EN, KI-04, KRYS-I, MGC and SANTINIS).We investigate the performance of several types of features (POS n-grams, character n-grams and word n-grams) and show that simple character n-grams perform best on current collections because of their ability to generalise both lexical and syntactic phenomena related to genres. However, we also show that these impressive results might not be transferrable to the wider web due to the lack of comparability between different annotation labels (many webpages cannot be described in terms of the genre labels in individual collections), lack of representativeness of existing collections (many genres are represented by webpages coming from a small number of sources) as well as problems in the reliability of genre annotation (many pages from the web are difficult to interpret in terms of the labels available). This suggests that more research is needed to understand genres on the Web.
We report on an on-going research project aimed at increasing the range of translation equivalents which can be automatically discovered by MT systems. The methodology is based on semi-supervised learning of indirect translation strategies from large comparable corpora and applying them in run-time to generate novel, previously unseen translation equivalents. This approach is different from methods based on parallel resources, which currently can reuse only individual translation equivalents. Instead it models translation strategies which generalise individual equivalents and can successfully generate an open class of new translation solutions. The task of the project is integration of the developed technology into open-source MT systems.
Cleaneval is a shared task and competitive evaluation on the topic of cleaning arbitrary web pages, with the goal of preparing web data for use as a corpus for linguistic and language technology research and development. The first exercise took place in 2007. We describe how it was set up, results, and lessons learnt
This paper reports the principles behind designing a tagset to cover Russian morphosyntactic phenomena, modifications of the core tagset, and its evaluation. The tagset is based on the MULTEXT-East framework, while the decisions in designing it were aimed at achieving a balance between parameters important for linguists and the possibility to detect and disambiguate them automatically. The final tagset contains about 500 tags and achieves about 95% accuracy on the disambiguated portion of the Russian National Corpus. We have also produced a test set that can be shared with other researchers.
This paper presents an approach to computer-assisted teaching of reading abilities using corpus data. The approach is supported by a set of tools for automatically selecting and classifying texts retrieved from the Internet. The approach is based on a linguistic model of textual cohesion which describes relations between larger textual units that go beyond the sentence level. We show that textual connectors that link such textual units reliably predict different types of texts, such as information and opinion: using only textual connectors as features, an SVM classifier achieves an F-score of between 0.85 and 0.93 for predicting these classes. The tools are used in our project on teaching reading skills in a cognate foreign language (L3) which is cognate to a known foreign language (L2).
In the paper we address two practical problems concerning the use of corpora in translation studies. The first stems from the limited resources available for targeted languages and genres within languages, whereas translation researchers and students need: sufficiently large modern corpora, either reflecting general language or specific to a problem domain. The second problem concerns the lackof a uniform interface for accessing the resources, even when the yexist. We deal with the first problem by developing a framework for semi-automatic acquisition of large corpora from the Internet for the languages relevant for our research and training needs. We outline the methodology used and discuss the composition of Internet-derived corpora. We deal with the second problem by developing a uniform interface to our corpora. In addition to standard options for choosingcorpora and sorting concordance lines, the interface can compute the list of collocations and filter the results according touser-specified patterns in order to detect language-specific syntacticstructures.
In this paper we present a tool for finding appropriate translation equivalents for words from the general lexicon using comparable corpora. For a phrase in the source language the tool suggests arange of possible expressions used in similar contexts in target language corpora. In the paper we discuss the method and present results of human evaluation of the performance of the tool.
In an age when demand for innovative and motivating language teaching methodologies is at a very high level, TREAT - the Trilingual REAding Tutor - combines the most advanced natural language processing (NLP) techniques with the latest second and third language acquisition (SLA/TLA) research in an intuitive and user-friendly environment that has been proven to help adult learners (native speakers of L1) acquire reading skills in an unknown L3 which is related to (cognate with) an L2 they know to some extent. This corpus-based methodology relies on existing linguistic resources, as well as materials that are easy to assemble, and can be adapted to support other pairs of L2-L3 related languages, as well. A small evaluation study conducted at the Leeds University Centre for Translation Studies indicates that, when using TREAT, learners feel more motivated to study an unknown L3, acquire significant linguistic knowledge of both the L3 and L2 rapidly, and increase their performance when translating from L3 into L1.