Natural Language Processing (NLP) research and development has experienced rapid progression in the recent times due to advances in deep learning. The introduction of pre-trained large language models (LLMs) is at the core of this transformation, significantly enhancing the performance of machine translation (MT) and speech technologies. This development has also led to fundamental changes in modern translation and speech tools and their methodologies. However, there remain challenges when extending this progress to underrepresented dialects and low-resource languages, primarily due to the need for more data. This paper details our submissions to the IWSLT speech translation (ST) tasks. We used the Whisper model for the automatic speech recognition (ASR) component. We then used mBART and NLLB as cascaded systems for utilising their MT capabilities. Our research primarily focused on exploring various dialects of low-resource languages and harnessing existing resources from linguistically related languages. We conducted our experiments for two morphologically diverse language pairs: Irish-to-English and Maltese-to-English. We used BLEU, chrF and COMET for evaluating our MT models.

This system description paper presents SETU-ADAPT’s submission to the WMT 2024 Biomedical Shared Task, where we participated for the language pairs English-to-French and English-to-German. Our approach focused on fine-tuning Large Language Models, using in-domain and synthetic data, employing different data augmentation and data retrieval strategies. We introduce a novel MT framework, involving three autonomous agents: a Translator Agent, an Evaluator Agent and a Reviewer Agent. We present our findings and report the quality of the outputs.

This paper presents the SETU-ADAPT’s submissions to the WMT 2024 Low-Resource Indic Language Translation task. We participated in the unconstrained segment of the task, focusing on the Assamese-to-English and English-to-Assamese language pairs. Our approach involves leveraging Large Language Models (LLMs) as the baseline systems for all our MT tasks. Furthermore, we applied various strategies to improve the baseline systems. In our first approach, we fine-tuned LLMs using all the data provided by the task organisers. Our second approach explores in-context learning by focusing on few-shot prompting. In our final approach we explore an efficient data extraction technique based on a fuzzy match-based similarity measure for fine-tuning. We evaluated our systems using BLEU, chrF, WER, and COMET. The experimental results showed that our strategies can effectively improve the quality of translations in low-resource scenarios.

This paper presents the SETU-ADAPT submissions to the WMT24 Chat Translation Task. Large language models (LLM) currently provides the state-of-the-art solutions in many natural language processing (NLP) problems including machine translation (MT). For the WMT24 Chat Translation Task we leveraged LLMs for their MT capabilities. In order to adapt the LLMs for a specific domain of interest, we explored different fine-tuning and prompting strategies. We also employed efficient data retrieval methods to curate the data used for fine-tuning. We carried out experiments for two language pairs: German-to-English and French-to-English. Our MT models were evaluated using three metrics: BLEU, chrF and COMET. In this paper we describes our experiments including training setups, results and findings.

Consistency is a key requirement of high-quality translation. It is especially important to adhere to pre-approved terminology and adapt to corrected translations in domain-specific projects. Machine translation (MT) has achieved significant progress in the area of domain adaptation. However, real-time adaptation remains challenging. Large-scale language models (LLMs) have recently shown interesting capabilities of in-context learning, where they learn to replicate certain input-output text generation patterns, without further fine-tuning. By feeding an LLM at inference time with a prompt that consists of a list of translation pairs, it can then simulate the domain and style characteristics. This work aims to investigate how we can utilize in-context learning to improve real-time adaptive MT. Our extensive experiments show promising results at translation time. For example, GPT-3.5 can adapt to a set of in-domain sentence pairs and/or terminology while translating a new sentence. We observe that the translation quality with few-shot in-context learning can surpass that of strong encoder-decoder MT systems, especially for high-resource languages. Moreover, we investigate whether we can combine MT from strong encoder-decoder models with fuzzy matches, which can further improve translation quality, especially for less supported languages. We conduct our experiments across five diverse language pairs, namely English-to-Arabic (EN-AR), English-to-Chinese (EN-ZH), English-to-French (EN-FR), English-to-Kinyarwanda (EN-RW), and English-to-Spanish (EN-ES).

This paper discusses the methods that we used for our submissions to the WMT 2023 Terminology Shared Task for German-to-English (DE-EN), English-to-Czech (EN-CS), and Chinese-to-English (ZH-EN) language pairs. The task aims to advance machine translation (MT) by challenging participants to develop systems that accurately translate technical terms, ultimately enhancing communication and understanding in specialised domains. To this end, we conduct experiments that utilise large language models (LLMs) for two purposes: generating synthetic bilingual terminology-based data, and post-editing translations generated by an MT model through incorporating pre-approved terms. Our system employs a four-step process: (i) using an LLM to generate bilingual synthetic data based on the provided terminology, (ii) fine-tuning a generic encoder-decoder MT model, with a mix of the terminology-based synthetic data generated in the first step and a randomly sampled portion of the original generic training data, (iii) generating translations with the fine-tuned MT model, and (iv) finally, leveraging an LLM for terminology-constrained automatic post-editing of the translations that do not include the required terms. The results demonstrate the effectiveness of our proposed approach in improving the integration of pre-approved terms into translations. The number of terms incorporated into the translations of the blind dataset increases from an average of 36.67% with the generic model to an average of 72.88% by the end of the process. In other words, successful utilisation of terms nearly doubles across the three language pairs.

Terms are essential indicators of a domain, and domain term translation is dealt with priority in any translation workflow. Translation service providers who use machine translation (MT) expect term translation to be unambiguous and consistent with the context and domain in question. Although current state-of-the-art neural MT (NMT) models are able to produce high-quality translations for many languages, they are still not at the level required when it comes to translating domain-specific terms. This study presents a terminology-aware instance- based adaptation method for improving terminology translation in NMT. We conducted our experiments for French-to-English and found that our proposed approach achieves a statistically significant improvement over the baseline NMT system in translating domain-specific terms. Specifically, the translation of multi-word terms is improved by 6.7% compared to the strong baseline.

Preservation of domain knowledge from the source to target is crucial in any translation workflow. It is common in the translation industry to receive highly-specialized projects, where there is hardly any parallel in-domain data. In such scenarios where there is insufficient in-domain data to fine-tune Machine Translation (MT) models, producing translations that are consistent with the relevant context is challenging. In this work, we propose leveraging state-of-the-art pretrained language models (LMs) for domain-specific data augmentation for MT, simulating the domain characteristics of either (a) a small bilingual dataset, or (b) the monolingual source text to be translated. Combining this idea with back-translation, we can generate huge amounts of synthetic bilingual in-domain data for both use cases. For our investigation, we used the state-of-the-art MT architecture, Transformer. We employed mixed fine-tuning to train models that significantly improve translation of in-domain texts. More specifically, our proposed methods achieved improvements of approximately 5-6 BLEU and 2-3 BLEU, respectively, on Arabic-to-English and English-to-Arabic language pairs. Furthermore, the outcome of human evaluation corroborates the automatic evaluation results.

Knowledge distillation (KD) can be used to reduce model size and training time, without significant loss in performance. However, the process of distilling knowledge requires translation of sizeable data sets, and the translation is usually performed using large cumbersome models (teacher models). Producing such translations for KD is expensive in terms of both time and cost, which is a significant concern for translation service providers. On top of that, this process can be the cause of higher carbon footprints. In this work, we tested different variants of a teacher model for KD, tracked the power consumption of the GPUs used during translation, recorded overall translation time, estimated translation cost, and measured the accuracy of the student models. The findings of our investigation demonstrate to the translation industry a cost-effective, high-quality alternative to the standard KD training methods.

Social media is not just meant for entertainment, it provides platforms for sharing information, news, facts and events. In the digital age, activists and numerous users are seen to be vocal regarding human rights and their violations in social media. However, their voices do not often reach to the targeted audience and concerned human rights organization. In this work, we aimed at detecting factual posts in social media about violation of human rights in any part of the world. The end product of this research can be seen as an useful asset for different peacekeeping organizations who could exploit it to monitor real-time circumstances about any incident in relation to violation of human rights. We chose one of the popular micro-blogging websites, Twitter, for our investigation. We used supervised learning algorithms in order to build human rights violation identification (HRVI) models which are able to identify Tweets in relation to incidents of human right violation. For this, we had to manually create a data set, which is one of the contributions of this research. We found that our classification models that were trained on this gold-standard dataset performed excellently in classifying factual Tweets about human rights violation, achieving an accuracy of upto 93% on hold-out test set.

Research on Machine Translation (MT) has achieved important breakthroughs in several areas. While there is much more to be done in order to build on this success, we believe that the language industry needs better ways to take full advantage of current achievements. Due to a combination of factors, including time, resources, and skills, businesses tend to apply pragmatism into their AI workflows. Hence, they concentrate more on outcomes, e.g. delivery, shipping, releases, and features, and adopt high-level working production solutions, where possible. Among the features thought to be helpful for translators are sentence-level and word-level translation auto-suggestion and auto-completion. Suggesting alternatives can inspire translators and limit their need to refer to external resources, which hopefully boosts their productivity. This work describes our submissions to WMT’s shared task on word-level auto-completion, for the Chinese-to-English, English-to-Chinese, German-to-English, and English-to-German language directions. We investigate the possibility of using pre-trained models and out-of-the-box features from available libraries. We employ random sampling to generate diverse alternatives, which reveals good results. Furthermore, we introduce our open-source API, based on CTranslate2, to serve translations, auto-suggestions, and auto-completions.

Emotion detection (ED) in tweets is a text classification problem that is of interest to Natural Language Processing (NLP) researchers. Code-mixing (CM) is a process of mixing linguistic units such as words of two different languages. The CM languages are characteristically different from the languages whose linguistic units are used for mixing. Whilst NLP has been shown to be successful for low-resource languages, it becomes challenging to perform NLP tasks on CM languages. As for ED, it has been rarely investigated on CM languages such as Hindi—English due to the lack of training data that is required for today’s data-driven classification algorithms. This research proposes a gold standard dataset for detecting emotions in CM Hindi–English tweets. This paper also presents our results about the investigation of the usefulness of our gold-standard dataset while testing a number of state-of-the-art classification algorithms. We found that the ED classifier built using SVM provided us the highest accuracy (75.17%) on the hold-out test set. This research would benefit the NLP community in detecting emotions from social media platforms in multilingual societies.

Interactive-predictive translation is a collaborative iterative process and where human translators produce translations with the help of machine translation (MT) systems interactively. Various sampling techniques in active learning (AL) exist to update the neural MT (NMT) model in the interactive-predictive scenario. In this paper and we explore term based (named entity count (NEC)) and quality based (quality estimation (QE) and sentence similarity (Sim)) sampling techniques – which are used to find the ideal candidates from the incoming data – for human supervision and MT model’s weight updation. We carried out experiments with three language pairs and viz. German-English and Spanish-English and Hindi-English. Our proposed sampling technique yields 1.82 and 0.77 and 0.81 BLEU points improvements for German-English and Spanish-English and Hindi-English and respectively and over random sampling based baseline. It also improves the present state-of-the-art by 0.35 and 0.12 BLEU points for German-English and Spanish-English and respectively. Human editing effort in terms of number-of-words-changed also improves by 5 and 4 points for German-English and Spanish-English and respectively and compared to the state-of-the-art.

This paper describes the ADAPT Centre’s submission to STAPLE (Simultaneous Translation and Paraphrase for Language Education) 2020, a shared task of the 4th Workshop on Neural Generation and Translation (WNGT), for the English-to-Portuguese translation task. In this shared task, the participants were asked to produce high-coverage sets of plausible translations given English prompts (input source sentences). We present our English-to-Portuguese machine translation (MT) models that were built applying various strategies, e.g. data and sentence selection, monolingual MT for generating alternative translations, and combining multiple n-best translations. Our experiments show that adding the aforementioned techniques to the baseline yields an excellent performance in the English-to-Portuguese translation task.

This paper describes the ADAPT Centre’s submissions to the WMT20 News translation shared task for English-to-Tamil and Tamil-to-English. We present our machine translation (MT) systems that were built using the state-of-the-art neural MT (NMT) model, Transformer. We applied various strategies in order to improve our baseline MT systems, e.g. onolin- gual sentence selection for creating synthetic training data, mining monolingual sentences for adapting our MT systems to the task, hyperparameters search for Transformer in lowresource scenarios. Our experiments show that adding the aforementioned techniques to the baseline yields an excellent performance in the English-to-Tamil and Tamil-to-English translation tasks.

This paper describes the ADAPT Centre’s submissions to the WMT20 Biomedical Translation Shared Task for English-to-Basque. We present the machine translation (MT) systems that were built to translate scientific abstracts and terms from biomedical terminologies, and using the state-of-the-art neural MT (NMT) model: Transformer. In order to improve our baseline NMT system, we employ a number of methods, e.g. “pseudo” parallel data selection, monolingual data selection for synthetic corpus creation, mining monolingual sentences for adapting our NMT systems to this task, hyperparameters search for Transformer in lowresource scenarios. Our experiments show that systematic addition of the aforementioned techniques to the baseline yields an excellent performance in the English-to-Basque translation task.

Statistical machine translation (SMT) which was the dominant paradigm in machine translation (MT) research for nearly three decades has recently been superseded by the end-to-end deep learning approaches to MT. Although deep neural models produce state-of-the-art results in many translation tasks, they are found to under-perform on resource-poor scenarios. Despite some success, none of the present-day benchmarks that have tried to overcome this problem can be regarded as a universal solution to the problem of translation of many low-resource languages. In this work, we investigate the performance of phrase-based SMT (PB-SMT) and neural MT (NMT) on a rarely-tested low-resource language-pair, English-to-Tamil, taking a specialised data domain (software localisation) into consideration. In particular, we produce rankings of our MT systems via a social media platform-based human evaluation scheme, and demonstrate our findings in the low-resource domain-specific text translation task.

Automatic recognition of customer complaints on products or services that they purchase can be crucial for the organisations, multinationals and online retailers since they can exploit this information to fulfil their customers’ expectations including managing and resolving the complaints. Recently, researchers have applied supervised learning strategies to automatically identify users’ complaints expressed in English on Twitter. The downside of these approaches is that they require labeled training data for learning, which is expensive to create. This poses a barrier for them being applied to low-resource languages and domains for which task-specific data is not available. Machine translation (MT) can be used as an alternative to the tools that require such task-specific data. In this work, we use state-of-the-art neural MT (NMT) models for translating Hindi reviews into English and investigate performance of the downstream classification task (complaints identification) on their English translations.

This paper describes the ADAPT Centre’s submission to the Adap-MT 2020 AI Translation Shared Task for English-to-Hindi. The neural machine translation (NMT) systems that we built to translate AI domain texts are state-of-the-art Transformer models. In order to improve the translation quality of our NMT systems, we made use of both in-domain and out-of-domain data for training and employed different fine-tuning techniques for adapting our NMT systems to this task, e.g. mixed fine-tuning and on-the-fly self-training. For this, we mined parallel sentence pairs and monolingual sentences from large out-of-domain data, and the mining process was facilitated through automatic extraction of terminology from the in-domain data. This paper outlines the experiments we carried out for this task and reports the performance of our NMT systems on the evaluation test set.

In interactive machine translation (MT), human translators correct errors in automatic translations in collaboration with the MT systems, which is seen as an effective way to improve the productivity gain in translation. In this study, we model source-language syntactic constituency parse and target-language syntactic descriptions in the form of supertags as conditional context for interactive prediction in neural MT (NMT). We found that the supertags significantly improve productivity gain in translation in interactive-predictive NMT (INMT), while syntactic parsing somewhat found to be effective in reducing human effort in translation. Furthermore, when we model this source- and target-language syntactic information together as the conditional context, both types complement each other and our fully syntax-informed INMT model statistically significantly reduces human efforts in a French–to–English translation task, achieving 4.30 points absolute (corresponding to 9.18% relative) improvement in terms of word prediction accuracy (WPA) and 4.84 points absolute (corresponding to 9.01% relative) reduction in terms of word stroke ratio (WSR) over the baseline.

This paper describes the ADAPT Centre sub-missions to WAT 2020 for the English-to-Odia translation task. We present the approaches that we followed to try to build competitive machine translation (MT) systems for English-to-Odia. Our approaches include monolingual data selection for creating synthetic data and identifying optimal sets of hyperparameters for the Transformer in a low-resource scenario. Our best MT system produces 4.96BLEU points on the evaluation test set in the English-to-Odia translation task.

In this paper we describe the ADAPT Centre’s submissions to the WAT 2020 document-level Business Scene Dialogue (BSD) Translation task. We only consider translating from Japanese to English for this task and we use the MarianNMT toolkit to train Transformer models. In order to improve the translation quality, we made use of both in-domain and out-of-domain data for training our Machine Translation (MT) systems, as well as various data augmentation techniques for fine-tuning the model parameters. This paper outlines the experiments we ran to train our systems and report the accuracy achieved through these various experiments.

Statistical machine translation (SMT) was the state-of-the-art in machine translation (MT) research for more than two decades, but has since been superseded by neural MT (NMT). Despite producing state-of-the-art results in many translation tasks, neural models underperform in resource-poor scenarios. Despite some success, none of the present-day benchmarks that have tried to overcome this problem can be regarded as a universal solution to the problem of translation of many low-resource languages. In this work, we investigate the performance of phrase-based SMT (PB-SMT) and NMT on two rarely-tested low-resource language-pairs, English-to-Tamil and Hindi-to-Tamil, taking a specialised data domain (software localisation) into consideration. This paper demonstrates our findings including the identification of several issues of the current neural approaches to low-resource domain-specific text translation.

Traditional statistical approaches to spelling correction usually consist of two consecutive processes — error detection and correction — and they are generally computationally intensive. Current state-of-the-art neural spelling correction models usually attempt to correct spelling errors directly over an entire sentence, which, as a consequence, lacks control of the process, e.g. they are prone to overcorrection. In recent years, recurrent neural networks (RNNs), in particular long short-term memory (LSTM) hidden units, have proven increasingly popular and powerful models for many natural language processing (NLP) problems. Accordingly, we made use of a bidirectional LSTM language model (LM) for our context-sensitive spelling detection and correction model which is shown to have much control over the correction process. While the use of LMs for spelling checking and correction is not new to this line of NLP research, our proposed approach makes better use of the rich neighbouring context, not only from before the word to be corrected, but also after it, via a dual-input deep LSTM network. Although in theory our proposed approach can be applied to any language, we carried out our experiments on Arabic, which we believe adds additional value given the fact that there are limited linguistic resources readily available in Arabic in comparison to many languages. Our experimental results demonstrate that the proposed methods are effective in both improving the quality of correction suggestions and minimising overcorrection.

Terminology translation plays a critical role in domain-specific machine translation (MT). In this paper, we conduct a comparative qualitative evaluation on terminology translation in phrase-based statistical MT (PB-SMT) and neural MT (NMT) in two translation directions: English-to-Hindi and Hindi-to-English. For this, we select a test set from a legal domain corpus and create a gold standard for evaluating terminology translation in MT. We also propose an error typology taking the terminology translation errors into consideration. We evaluate the MT systems’ performance on terminology translation, and demonstrate our findings, unraveling strengths, weaknesses, and similarities of PB-SMT and NMT in the area of term translation.

This paper presents a case-study of work done by Applied Language Solutions (ALS) for a large social networking provider who claim to have built the world’s first multi-language social network, where Internet users from all over the world can communicate in languages that are available in the system. In an initial phase, the social networking provider contracted ALS to build Machine Translation (MT) engines for twelve language-pairs: Russian⇔English, Russian⇔Turkish, Russian⇔Arabic, Turkish⇔English, Turkish⇔Arabic and Arabic⇔English. All of the input data is user-generated content, so we faced a number of problems in building large-scale, robust, high-quality engines. Primarily, much of the source-language data is of ‘poor’ or at least ‘non-standard’ quality. This comes in many forms: (i) content produced by non-native speakers, (ii) content produced by native speakers containing non-deliberate typos, or (iii) content produced by native speakers which deliberately departs from spelling norms to bring about some linguistic effect. Accordingly, in addition to the ‘regular’ pre-processing techniques used in the building of our statistical MT systems, we needed to develop routines to deal with all these scenarios. In this paper, we describe how we handle shortforms, acronyms, typos, punctuation errors, non-dictionary slang, wordplay, censor avoidance and emoticons. We demonstrate automatic evaluation scores on the social network data, together with insights from the the social networking provider regarding some of the typical errors made by the MT engines, and how we managed to correct these in the engines.

Content localisation via machine translation (MT) is a sine qua non, especially for international online business. While most applications utilise rule-based solutions due to the lack of suitable in-domain parallel corpora for statistical MT (SMT) training, in this paper we investigate the possibility of applying SMT where huge amounts of monolingual content only are available. We describe a case study where an analysis of a very large amount of monolingual online trading data from eBay is conducted by ALS with a view to reducing this corpus to the most representative sample in order to ensure the widest possible coverage of the total data set. Furthermore, minimal yet optimal sets of sentences/words/terms are selected for generation of initial translation units for future SMT system-building.

Statistical machine translation (SMT) models have recently begun to include source context modeling, under the assumption that the proper lexical choice of the translation for an ambiguous word can be determined from the context in which it appears. Various types of lexical and syntactic features have been explored as effective source context to improve phrase selection in SMT. In the present work, we introduce lexico-syntactic descriptions in the form of supertags as source-side context features in the state-of-the-art hierarchical phrase-based SMT (HPB) model. These features enable us to exploit source similarity in addition to target similarity, as modelled by the language model. In our experiments two kinds of supertags are employed: those from lexicalized tree-adjoining grammar (LTAG) and combinatory categorial grammar (CCG). We use a memory-based classification framework that enables the efficient estimation of these features. Despite the differences between the two supertagging approaches, they give similar improvements. We evaluate the performance of our approach on an English-to-Dutch translation task, and report statistically significant improvements of 4.48% and 6.3% BLEU scores in translation quality when adding CCG and LTAG supertags, respectively, as context-informed features.