Structured prediction tasks, like machine translation, involve learning functions that map structured inputs to structured outputs. Recurrent Neural Networks (RNNs) have historically been a popular choice for such tasks, including in natural language processing (NLP) applications. However, training RNNs using Maximum Likelihood Estimation (MLE) has its limitations, including exposure bias and a mismatch between training and testing metrics. SEARNN, based on the learning to search (L2S) framework, has been proposed as an alternative to MLE for RNN training. This project explored the potential of SEARNN to improve machine translation for low-resourced African languages -- a challenging task characterized by limited training data availability and the morphological complexity of the languages. Through experiments conducted on translation for English to Igbo, French to \ewe, and French to \ghomala directions, this project evaluated the efficacy of SEARNN over MLE in addressing the unique challenges posed by these languages. With an average BLEU score improvement of $5.4$\% over the MLE objective, we proved that SEARNN is indeed a viable algorithm to effectively train RNNs on machine translation for low-resourced languages.

Recent advancements in machine translation (MT) have significantly enhanced translation quality across various domains. However, the translation of literary texts remains a formidable challenge due to their complex language, figurative expressions, and cultural nuances. In this work, we introduce a novel multi-agent framework based on large language models (LLMs) for literary translation, implemented as a company called TransAgents, which mirrors traditional translation publication process by leveraging the collective capabilities of multiple agents, to address the intricate demands of translating literary works. To evaluate the effectiveness of our system, we propose two innovative evaluation strategies: Monolingual Human Preference (MHP) and Bilingual LLM Preference (BLP). MHP assesses translations from the perspective of monolingual readers of the target language, while BLP uses advanced LLMs to compare translations directly with the original texts. Empirical findings indicate that despite lower d-BLEU scores, translations from TransAgents are preferred by both human evaluators and LLMs over human-written references, particularly in genres requiring domain-specific knowledge. We also highlight the strengths and limitations of TransAgents through case studies and suggests directions for future research.

People use language for various purposes. Apart from sharing information, individuals may use it to express emotions or to show respect for another person. In this paper, we focus on the formality level of machine-generated translations and present FAME-MT -- a dataset consisting of 11.2 million translations between 15 European source languages and 8 European target languages classified to formal and informal classes according to target sentence formality. This dataset can be used to fine-tune machine translation models to ensure a given formality level for each European target language considered. We describe the dataset creation procedure, the analysis of the dataset's quality showing that FAME-MT is a reliable source of language register information, and we present a publicly available proof-of-concept machine translation model that uses the dataset to steer the formality level of the translation. Currently, it is the largest dataset of formality annotations, with examples expressed in 112 European language pairs. The dataset is published online: https://github.com/laniqo-public/fame-mt/ .

With the advent of Neural Machine Translation (NMT) systems, the MT output has reached unprecedented accuracy levels which resulted in the ubiquity of MT tools on almost all online platforms with multilingual content. However, NMT systems, like other state-of-the-art AI generative systems, are prone to errors that are deemed machine hallucinations. The problem with NMT hallucinations is that they are remarkably \textit{fluent} hallucinations. Since they are trained to produce grammatically correct utterances, NMT systems are capable of producing mistranslations that are too fluent to be recognised by both users of the MT tool, as well as by automatic quality metrics that are used to gauge their performance. In this paper, we introduce an authentic dataset of machine translation critical errors to point to the ethical and safety issues involved in the common use of MT. The dataset comprises mistranslations of Arabic mental health postings manually annotated with critical error types. We also show how the commonly used quality metrics do not penalise critical errors and highlight this as a critical issue that merits further attention from researchers.

Lexicon or dictionary generation across domains is of significant societal importance, as it can potentially enhance information accessibility for a diverse user base while preserving language identity. Prior work in the field primarily focuses on bilingual lexical induction, which deals with word alignments using mapping-based or corpora-based approaches. Though initiated by researchers, the research associated with lexicon generation is limited, even more so with domain-specific lexicons. This task becomes particularly important in atypical medical, engineering, and other technical domains, owing to the highly infrequent usage of the terms and negligibly low data availability of technical terms in many low-resource languages. Owing to the research gap in lexicon generation, especially with a limited focus on the domain-specific area, we propose a new model to generate dictionary words for 6 Indian languages in the multi-domain setting. Our model consists of domain-specific and domain-generic layers that encode information, and these layers are invoked via a learnable routing technique. Further, we propose an approach to explicitly leverage the relatedness between these Indian languages toward coherent translation. We also release a new benchmark dataset across 6 Indian languages that span 8 diverse domains that can propel further research in domain-specific lexicon induction. We conduct both zero-shot and few-shot experiments across multiple domains to show the efficacy of our proposed model in generalizing to unseen domains and unseen languages.

Subtitling plays a crucial role in enhancing the accessibility of audiovisual content and encompasses three primary subtasks: translating spoken dialogue, segmenting translations into concise textual units, and estimating timestamps that govern their on-screen duration. Past attempts to automate this process rely, to varying degrees, on automatic transcripts, employed diversely for the three subtasks. In response to the acknowledged limitations associated with this reliance on transcripts, recent research has shifted towards transcription-free solutions for translation and segmentation, leaving the direct generation of timestamps as uncharted territory. To fill this gap, we introduce the first direct model capable of producing automatic subtitles, entirely eliminating any dependence on intermediate transcripts also for timestamp prediction. Experimental results, backed by manual evaluation, showcase our solution's new state-of-the-art performance across multiple language pairs and diverse conditions.

In this paper, we introduce the first comprehensive multilingual sign language dataset named Prompt2Sign, which builds from public data including American Sign Language (ASL) and seven others. Our dataset transforms a vast array of videos into a streamlined, model-friendly format, optimized for training with translation models like seq2seq and text2text. Building on this new dataset, we propose SignLLM, the first multilingual Sign Language Production (SLP) model, which includes two novel multilingual SLP modes that allow for the generation of sign language gestures from input text or prompt. Both of the modes can use a new loss and a module based on reinforcement learning, which accelerates the training by enhancing the model's capability to autonomously sample high-quality data. We present benchmark results of SignLLM, which demonstrate that our model achieves state-of-the-art performance on SLP tasks across eight sign languages. More code and materials are available at https://signllm.github.io/.

Toxicity mitigation consists in rephrasing text in order to remove offensive or harmful meaning. Neural natural language processing (NLP) models have been widely used to target and mitigate textual toxicity. However, existing methods fail to detoxify text while preserving the initial non-toxic meaning at the same time. In this work, we propose to apply counterfactual generation methods from the eXplainable AI (XAI) field to target and mitigate textual toxicity. In particular, we perform text detoxification by applying local feature importance and counterfactual generation methods to a toxicity classifier distinguishing between toxic and non-toxic texts. We carry out text detoxification through counterfactual generation on three datasets and compare our approach to three competitors. Automatic and human evaluations show that recently developed NLP counterfactual generators can mitigate toxicity accurately while better preserving the meaning of the initial text as compared to classical detoxification methods. Finally, we take a step back from using automated detoxification tools, and discuss how to manage the polysemous nature of toxicity and the risk of malicious use of detoxification tools. This work is the first to bridge the gap between counterfactual generation and text detoxification and paves the way towards more practical application of XAI methods.

Transliterating related languages that use different scripts into a common script shows effectiveness in improving crosslingual transfer in downstream tasks. However, this methodology often makes pretraining a model from scratch unavoidable, as transliteration brings about new subwords not covered in existing multilingual pretrained language models (mPLMs). This is not desired because it takes a lot of computation budget for pretraining. A more promising way is to make full use of available mPLMs. To this end, this paper proposes a simple but effective framework: Transliterate-Merge-Initialize (TransMI), which can create a strong baseline well-suited for data that is transliterated into a common script by exploiting an mPLM and its accompanied tokenizer. TransMI has three stages: (a) transliterate the vocabulary of an mPLM into a common script; (b) merge the new vocabulary with the original vocabulary; and (c) initialize the embeddings of the new subwords. We applied TransMI to three recent strong mPLMs, and our experiments demonstrate that TransMI not only preserves their ability to handle non-transliterated data, but also enables the models to effectively process transliterated data: the results show a consistent improvement of 3% to 34%, varying across different models and tasks. We make our code and models publicly available at \url{https://github.com/cisnlp/TransMI}.

Authorship obfuscation techniques hold the promise of helping people protect their privacy in online communications by automatically rewriting text to hide the identity of the original author. However, obfuscation has been evaluated in narrow settings in the NLP literature and has primarily been addressed with superficial edit operations that can lead to unnatural outputs. In this work, we introduce an automatic text privatization framework that fine-tunes a large language model via reinforcement learning to produce rewrites that balance soundness, sense, and privacy. We evaluate it extensively on a large-scale test set of English Reddit posts by 68k authors composed of short-medium length texts. We study how the performance changes among evaluative conditions including authorial profile length and authorship detection strategy. Our method maintains high text quality according to both automated metrics and human evaluation, and successfully evades several automated authorship attacks.