Machine Translation (MT) evaluation metrics assess translation quality automatically. Recently, researchers have employed MT metrics for various new use cases, such as data filtering and translation re-ranking. However, most MT metrics return assessments as scalar scores that are difficult to interpret, posing a challenge to making informed design choices. Moreover, MT metrics' capabilities have historically been evaluated using correlation with human judgment, which, despite its efficacy, falls short of providing intuitive insights into metric performance, especially in terms of new metric use cases. To address these issues, we introduce an interpretable evaluation framework for MT metrics. Within this framework, we evaluate metrics in two scenarios that serve as proxies for the data filtering and translation re-ranking use cases. Furthermore, by measuring the performance of MT metrics using Precision, Recall, and F-score, we offer clearer insights into their capabilities than correlation with human judgments. Finally, we raise concerns regarding the reliability of manually curated data following the Direct Assessments+Scalar Quality Metrics (DA+SQM) guidelines, reporting a notably low agreement with Multidimensional Quality Metrics (MQM) annotations.
Models for streaming speech translation (ST) can achieve high accuracy and low latency if they're developed with vast amounts of paired audio in the source language and written text in the target language. Yet, these text labels for the target language are often pseudo labels due to the prohibitive cost of manual ST data labeling. In this paper, we introduce a methodology named Connectionist Temporal Classification guided modality matching (CTC-GMM) that enhances the streaming ST model by leveraging extensive machine translation (MT) text data. This technique employs CTC to compress the speech sequence into a compact embedding sequence that matches the corresponding text sequence, allowing us to utilize matched {source-target} language text pairs from the MT corpora to refine the streaming ST model further. Our evaluations with FLEURS and CoVoST2 show that the CTC-GMM approach can increase translation accuracy relatively by 13.9% and 6.4% respectively, while also boosting decoding speed by 59.7% on GPU.
Grammar induction has made significant progress in recent years. However, it is not clear how the application of induced grammar could enhance practical performance in downstream tasks. In this work, we introduce an unsupervised grammar induction method for language understanding and generation. We construct a grammar parser to induce constituency structures and dependency relations, which is simultaneously trained on downstream tasks without additional syntax annotations. The induced grammar features are subsequently incorporated into Transformer as a syntactic mask to guide self-attention. We evaluate and apply our method to multiple machine translation tasks and natural language understanding tasks. Our method demonstrates superior performance compared to the original Transformer and other models enhanced with external parsers. Experimental results indicate that our method is effective in both from-scratch and pre-trained scenarios. Additionally, our research highlights the contribution of explicitly modeling the grammatical structure of texts to neural network models.
Leveraging large language models (LLMs) for various natural language processing tasks has led to superlative claims about their performance. For the evaluation of machine translation (MT), existing research shows that LLMs are able to achieve results comparable to fine-tuned multilingual pre-trained language models. In this paper, we explore what translation information, such as the source, reference, translation errors and annotation guidelines, is needed for LLMs to evaluate MT quality. In addition, we investigate prompting techniques such as zero-shot, Chain of Thought (CoT) and few-shot prompting for eight language pairs covering high-, medium- and low-resource languages, leveraging varying LLM variants. Our findings indicate the importance of reference translations for an LLM-based evaluation. While larger models do not necessarily fare better, they tend to benefit more from CoT prompting, than smaller models. We also observe that LLMs do not always provide a numerical score when generating evaluations, which poses a question on their reliability for the task. Our work presents a comprehensive analysis for resource-constrained and training-less LLM-based evaluation of machine translation. We release the accrued prompt templates, code and data publicly for reproducibility.
Machine translation (MT) of user-generated content (UGC) poses unique challenges, including handling slang, emotion, and literary devices like irony and sarcasm. Evaluating the quality of these translations is challenging as current metrics do not focus on these ubiquitous features of UGC. To address this issue, we utilize an existing emotion-related dataset that includes emotion labels and human-annotated translation errors based on Multi-dimensional Quality Metrics. We extend it with sentence-level evaluation scores and word-level labels, leading to a dataset suitable for sentence- and word-level translation evaluation and emotion classification, in a multi-task setting. We propose a new architecture to perform these tasks concurrently, with a novel combined loss function, which integrates different loss heuristics, like the Nash and Aligned losses. Our evaluation compares existing fine-tuning and multi-task learning approaches, assessing generalization with ablative experiments over multiple datasets. Our approach achieves state-of-the-art performance and we present a comprehensive analysis for MT evaluation of UGC.
Large language models (LLMs) have achieved remarkable success across various NLP tasks, yet their focus has predominantly been on English due to English-centric pre-training and limited multilingual data. While some multilingual LLMs claim to support for hundreds of languages, models often fail to provide high-quality response for mid- and low-resource languages, leading to imbalanced performance heavily skewed in favor of high-resource languages like English and Chinese. In this paper, we prioritize quality over scaling number of languages, with a focus on multilingual machine translation task, and introduce X-ALMA, a model designed with a commitment to ensuring top-tier performance across 50 diverse languages, regardless of their resource levels. X-ALMA surpasses state-of-the-art open-source multilingual LLMs, such as Aya-101 and Aya-23, in every single translation direction on the FLORES and WMT'23 test datasets according to COMET-22. This is achieved by plug-and-play language-specific module architecture to prevent language conflicts during training and a carefully designed training regimen with novel optimization methods to maximize the translation performance. At the final stage of training regimen, our proposed Adaptive Rejection Preference Optimization (ARPO) surpasses existing preference optimization methods in translation tasks.
Achieving consistent high-quality machine translation (MT) across diverse domains remains a significant challenge, primarily due to the limited and imbalanced parallel training data available in various domains. While large language models (LLMs) have demonstrated impressive general understanding and generation abilities, their potential in multi-domain MT is under-explored. We establish a comprehensive benchmark for multi-domain translation, featuring 25 German$\Leftrightarrow$English and 22 Chinese$\Leftrightarrow$English test sets respectively covering 15 domains. Our evaluation of prominent LLMs reveals a discernible performance gap against traditional MT systems, highlighting domain overfitting and catastrophic forgetting issues after fine-tuning on domain-limited corpora. To mitigate this, we propose a domain Chain of Thought (CoT) fine-tuning technique that utilizes the intrinsic multi-domain intelligence of LLMs to improve translation performance. This method inspires the LLM to perceive domain information from the source text, which then serves as a helpful hint to guide the translation process. Despite being trained on a small dataset of four domains, our CoT fine-tune approach achieves notable enhancements in translation accuracy and domain robustness than traditional fine-tuning, as evidenced by an average 1.53 BLEU score increase in over 20 German$\rightarrow$English distinct out-of-domain tests.
Preference Optimization (PO) techniques are currently one of the state of the art techniques for fine-tuning large language models (LLMs) on pairwise preference feedback from human annotators. However, in machine translation, this sort of feedback can be difficult to solicit. Additionally, Kreutzer et al. (2018) have shown that, for machine translation, pairwise preferences are less reliable than other forms of human feedback, such as 5-point ratings. We examine post-edits to see if they can be a source of reliable human preferences by construction. In PO, a human annotator is shown sequences $s_1$ and $s_2$ and asked for a preference judgment, %$s_1 > s_2$; while for post-editing, editors create $s_1$ and know that it should be better than $s_2$. We attempt to use these implicit preferences for PO and show that it helps the model move towards post-edit-like hypotheses and away from machine translation-like hypotheses. Furthermore, we show that best results are obtained by pre-training the model with supervised fine-tuning (SFT) on post-edits in order to promote post-edit-like hypotheses to the top output ranks.
This paper investigates the impact of verbose LLM translations on evaluation. We first demonstrate the prevalence of this behavior across several LLM outputs drawn from the WMT 2024 general shared task on machine translation. We then identify the primary triggers of verbosity, including safety, copyright concerns, and insufficient context in short input queries. Finally, we show that ignoring this behavior unfairly penalizes more verbose LLMs according to both automatic and human evaluations, highlighting the need to address this issue for more accurate future evaluations.
Gender bias in machine translation (MT) is recognized as an issue that can harm people and society. And yet, advancements in the field rarely involve people, the final MT users, or inform how they might be impacted by biased technologies. Current evaluations are often restricted to automatic methods, which offer an opaque estimate of what the downstream impact of gender disparities might be. We conduct an extensive human-centered study to examine if and to what extent bias in MT brings harms with tangible costs, such as quality of service gaps across women and men. To this aim, we collect behavioral data from 90 participants, who post-edited MT outputs to ensure correct gender translation. Across multiple datasets, languages, and types of users, our study shows that feminine post-editing demands significantly more technical and temporal effort, also corresponding to higher financial costs. Existing bias measurements, however, fail to reflect the found disparities. Our findings advocate for human-centered approaches that can inform the societal impact of bias.
This paper addresses the challenge of accurately translating technical terms, which are crucial for clear communication in specialized fields. We introduce the Parenthetical Terminology Translation (PTT) task, designed to mitigate potential inaccuracies by displaying the original term in parentheses alongside its translation. To implement this approach, we generated a representative PTT dataset using a collaborative approach with large language models and applied knowledge distillation to fine-tune traditional Neural Machine Translation (NMT) models and small-sized Large Language Models (sLMs). Additionally, we developed a novel evaluation metric to assess both overall translation accuracy and the correct parenthetical presentation of terms. Our findings indicate that sLMs did not consistently outperform NMT models, with fine-tuning proving more effective than few-shot prompting, particularly in models with continued pre-training in the target language. These insights contribute to the advancement of more reliable terminology translation methodologies.
Recent advancements in large language models (LLMs) have shown promising results in multilingual translation even with limited bilingual supervision. The major challenges are catastrophic forgetting and parameter interference for finetuning LLMs when provided parallel training data. To address these challenges, we propose LANDeRMT, a \textbf{L}anguage-\textbf{A}ware \textbf{N}euron \textbf{De}tecting and \textbf{R}outing framework that selectively finetunes LLMs to \textbf{M}achine \textbf{T}ranslation with diverse translation training data. In LANDeRMT, we evaluate the awareness of neurons to MT tasks and categorize them into language-general and language-specific neurons. This categorization enables selective parameter updates during finetuning, mitigating parameter interference and catastrophic forgetting issues. For the detected neurons, we further propose a conditional awareness-based routing mechanism to dynamically adjust language-general and language-specific capacity within LLMs, guided by translation signals. Experimental results demonstrate that the proposed LANDeRMT is very effective in learning translation knowledge, significantly improving translation quality over various strong baselines for multiple language pairs.
Extremely low-resource (XLR) languages lack substantial corpora for training NLP models, motivating the use of all available resources such as dictionaries and grammar books. Machine Translation from One Book (Tanzer et al., 2024) suggests prompting long-context LLMs with one grammar book enables English-Kalamang translation, an unseen XLR language - a noteworthy case of linguistic knowledge helping an NLP task. We investigate whether the book's grammatical explanations or its parallel examples are most effective for learning XLR translation, finding almost all improvement stems from the parallel examples. Further, we find similar results for Nepali, a seen low-resource language, and achieve performance comparable to an LLM with a grammar book by simply fine-tuning an encoder-decoder translation model. We then investigate where grammar books help by testing two linguistic tasks, grammaticality judgment and gloss prediction, and we explore what kind of grammatical knowledge helps by introducing a typological feature prompt that achieves leading results on these more relevant tasks. We thus emphasise the importance of task-appropriate data for XLR languages: parallel examples for translation, and grammatical data for linguistic tasks. As we find no evidence that long-context LLMs can make effective use of grammatical explanations for XLR translation, we suggest data collection for multilingual XLR tasks such as translation is best focused on parallel data over linguistic description.
Singlish, or formally Colloquial Singapore English, is an English-based creole language originating from the SouthEast Asian country Singapore. The language contains influences from Sinitic languages such as Chinese dialects, Malay, Tamil and so forth. A fundamental task to understanding Singlish is to first understand the pragmatic functions of its discourse particles, upon which Singlish relies heavily to convey meaning. This work offers a preliminary effort to disentangle the Singlish discourse particles (lah, meh and hor) with task-driven representation learning. After disentanglement, we cluster these discourse particles to differentiate their pragmatic functions, and perform Singlish-to-English machine translation. Our work provides a computational method to understanding Singlish discourse particles, and opens avenues towards a deeper comprehension of the language and its usage.
Neural metrics for machine translation (MT) evaluation have become increasingly prominent due to their superior correlation with human judgments compared to traditional lexical metrics. Researchers have therefore utilized neural metrics through quality-informed decoding strategies, achieving better results than likelihood-based methods. With the rise of Large Language Models (LLMs), preference-based alignment techniques have gained attention for their potential to enhance translation quality by optimizing model weights directly on preferences induced by quality estimators. This study focuses on Contrastive Preference Optimization (CPO) and conducts extensive experiments to evaluate the impact of preference-based alignment on translation quality. Our findings indicate that while CPO consistently outperforms Supervised Fine-Tuning (SFT) on high-quality data with regard to the alignment metric, it may lead to instability across downstream evaluation metrics, particularly between neural and lexical ones. Additionally, we demonstrate that relying solely on the base model for generating candidate translations achieves performance comparable to using multiple external systems, while ensuring better consistency across downstream metrics.
Cross-lingual cross-modal retrieval (CCR) aims to retrieve visually relevant content based on non-English queries, without relying on human-labeled cross-modal data pairs during training. One popular approach involves utilizing machine translation (MT) to create pseudo-parallel data pairs, establishing correspondence between visual and non-English textual data. However, aligning their representations poses challenges due to the significant semantic gap between vision and text, as well as the lower quality of non-English representations caused by pre-trained encoders and data noise. To overcome these challenges, we propose LECCR, a novel solution that incorporates the multi-modal large language model (MLLM) to improve the alignment between visual and non-English representations. Specifically, we first employ MLLM to generate detailed visual content descriptions and aggregate them into multi-view semantic slots that encapsulate different semantics. Then, we take these semantic slots as internal features and leverage them to interact with the visual features. By doing so, we enhance the semantic information within the visual features, narrowing the semantic gap between modalities and generating local visual semantics for subsequent multi-level matching. Additionally, to further enhance the alignment between visual and non-English features, we introduce softened matching under English guidance. This approach provides more comprehensive and reliable inter-modal correspondences between visual and non-English features. Extensive experiments on four CCR benchmarks, \ie Multi30K, MSCOCO, VATEX, and MSR-VTT-CN, demonstrate the effectiveness of our proposed method. Code: \url{https://github.com/LiJiaBei-7/leccr}.
For crosslingual conversation and trade, Neural Machine Translation (NMT) is pivotal yet faces persistent challenges with monotony and repetition in generated content. Traditional solutions that rely on penalizing text redundancy or token reoccurrence have shown limited efficacy, particularly for lengthy article and e-commerce descriptions with inherent redundancy, even with the advent of Large Language Models (LLMs). This paper investigates the underlying causes of textual repetition through the lens of information entropy, attributing the phenomenon to the elevated uncertainty within the input text. To address this, a novel algorithm named Contrastive Token Learning with Similarity Decay (CTSD) is introduced, which modulates the suppression of tokens dynamically, informed by varying attention weights and inter-token distances. Furthermore, an e-commerce dataset comprised of title texts of online real items is compiled and released susceptible to hallucination translations to benchmark the algorithm. Extensive evaluations demonstrate that CTSD significantly outperforms existing approaches in precision and generalizability. Additional online A/B testing underscores its practical value, showing marked improvements in user engagement and conversion. Notably, this method has been implemented with full traffic on eight multilingual sites of alibaba.com, the largest B2B e-commerce platform in the world.
We investigate zero-shot cross-lingual news sentiment detection, aiming to develop robust sentiment classifiers that can be deployed across multiple languages without target-language training data. We introduce novel evaluation datasets in several less-resourced languages, and experiment with a range of approaches including the use of machine translation; in-context learning with large language models; and various intermediate training regimes including a novel task objective, POA, that leverages paragraph-level information. Our results demonstrate significant improvements over the state of the art, with in-context learning generally giving the best performance, but with the novel POA approach giving a competitive alternative with much lower computational overhead. We also show that language similarity is not in itself sufficient for predicting the success of cross-lingual transfer, but that similarity in semantic content and structure can be equally important.
Reinforcement Learning from Human Feedback (RLHF) and derivative techniques like Direct Preference Optimization (DPO) are task-alignment algorithms used to repurpose general, foundational models for specific tasks. We show that applying task-alignment to neural machine translation (NMT) addresses an existing task--data mismatch in NMT, leading to improvements across all languages of a multilingual model, even when task-alignment is only applied to a subset of those languages. We do so by introducing Direct Quality Optimization (DQO), a variant of DPO leveraging a pre-trained translation quality estimation model as a proxy for human preferences, and verify the improvements with both automatic metrics and human evaluation.
We derive and investigate two DPO variants that explicitly model the possibility of declaring a tie in pair-wise comparisons. We replace the Bradley-Terry model in DPO with two well-known modeling extensions, by Rao and Kupper and by Davidson, that assign probability to ties as alternatives to clear preferences. Our experiments in neural machine translation and summarization show that explicitly labeled ties can be added to the datasets for these DPO variants without the degradation in task performance that is observed when the same tied pairs are presented to DPO. We find empirically that the inclusion of ties leads to stronger regularization with respect to the reference policy as measured by KL divergence, and we see this even for DPO in its original form. These findings motivate and enable the inclusion of tied pairs in preference optimization as opposed to simply discarding them.