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.

This paper presents a dataset for evaluating the machine translation of emotion-loaded user generated content. It contains human-annotated quality evaluation data and post-edited reference translations. The dataset is available at our GitHub repository.

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.

This paper investigates whether large language models (LLMs) are state-of-the-art quality estimators for machine translation of user-generated content (UGC) that contains emotional expressions, without the use of reference translations. To achieve this, we employ an existing emotion-related dataset with human-annotated errors and calculate quality evaluation scores based on the Multi-dimensional Quality Metrics. We compare the accuracy of several LLMs with that of our fine-tuned baseline models, under in-context learning and parameter-efficient fine-tuning (PEFT) scenarios. We find that PEFT of LLMs leads to better performance in score prediction with human interpretable explanations than fine-tuned models. However, a manual analysis of LLM outputs reveals that they still have problems such as refusal to reply to a prompt and unstable output while evaluating machine translation of UGC.

Abbreviations and their associated long forms are important textual elements that are present in almost every scientific communication, and having information about these forms can help improve several NLP tasks. In this paper, our aim is to fine-tune language models for automatically identifying abbreviations and long forms. We used existing datasets which are annotated with abbreviations and long forms to train and test several language models, including transformer models, character-level language models, stacking of different embeddings, and ensemble methods. Our experiments showed that it was possible to achieve state-of-the-art results by stacking RoBERTa embeddings with domain-specific embeddings. However, the analysis of our first run showed that one of the datasets had issues in the BIO annotation, which led us to propose a revised dataset. After re-training selected models on the revised dataset, results show that character-level models achieve comparable results, especially when detecting abbreviations, but both RoBERTa large and the stacking of embeddings presented better results on biomedical data. When tested on a different subdomain (segments extracted from computer science texts), an ensemble method proved to yield the best results for the detection of long forms, and a character-level model had the best performance in detecting abbreviations.

In this paper, we focus on how current Machine Translation (MT) engines perform on the translation of emotion-loaded texts by evaluating outputs from Google Translate according to a framework proposed in this paper. We propose this evaluation framework based on the Multidimensional Quality Metrics (MQM) and perform detailed error analyses of the MT outputs. From our analysis, we observe that about 50% of MT outputs are erroneous in preserving emotions. After further analysis of the erroneous examples, we find that emotion carrying words and linguistic phenomena such as polysemous words, negation, abbreviation etc., are common causes for these translation errors.

This paper attempts to identify challenges professional translators face when translating emotion-loaded texts as well as errors machine translation (MT) makes when translating this content. We invited ten Chinese-English translators to translate thirty posts of a Chinese microblog, and interviewed them about the challenges encountered during translation and the problems they believe MT might have. Further, we analysed more than five-thousand automatic translations of microblog posts to observe problems in MT outputs. We establish that the most challenging problem for human translators is emotion-carrying words, which translators also consider as a problem for MT. Analysis of MT outputs shows that this is also the most common source of MT errors. We also find that what is challenging for MT, such as non-standard writing, is not necessarily an issue for humans. Our work contributes to a better understanding of the challenges for the translation of microblog posts by humans and MT, caused by different forms of expression of emotion.

This paper summarises the submissions our team, SURREY-CTS-NLP has made for the WASSA 2022 Shared Task for the prediction of empathy, distress and emotion. In this work, we tested different learning strategies, like ensemble learning and multi-task learning, as well as several large language models, but our primary focus was on analysing and extracting emotion-intensive features from both the essays in the training data and the news articles, to better predict empathy and distress scores from the perspective of discourse and sentiment analysis. We propose several text feature extraction schemes to compensate the small size of training examples for fine-tuning pretrained language models, including methods based on Rhetorical Structure Theory (RST) parsing, cosine similarity and sentiment score. Our best submissions achieve an average Pearson correlation score of 0.518 for the empathy prediction task and an F1 score of 0.571 for the emotion prediction task, indicating that using these schemes to extract emotion-intensive information can help improve model performance.