LAYERED: Metric for Machine Translation Evaluation

Proceedings of the Second Workshop on …, 2007

Meteor is an automatic metric for Machine Translation evaluation which has been demonstrated to have high levels of correlation with human judgments of translation quality, significantly outperforming the more commonly used Bleu metric. It is one of several automatic metrics used in this year's shared task within the ACL WMT-07 workshop. This paper recaps the technical details underlying the metric and describes recent improvements in the metric. The latest release includes improved metric parameters and extends the metric to support evaluation of MT output in Spanish, French and German, in addition to English.

Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP: Volume 1 - ACL-IJCNLP '09, 2009

A number of approaches to Automatic MT Evaluation based on deep linguistic knowledge have been suggested. However, n-gram based metrics are still today the dominant approach. The main reason is that the advantages of employing deeper linguistic information have not been clarified yet. In this work, we propose a novel approach for meta-evaluation of MT evaluation metrics, since correlation cofficient against human judges do not reveal details about the advantages and disadvantages of particular metrics. We then use this approach to investigate the benefits of introducing linguistic features into evaluation metrics. Overall, our experiments show that (i) both lexical and linguistic metrics present complementary advantages and (ii) combining both kinds of metrics yields the most robust metaevaluation performance.

Language Resources and Evaluation, 2018

Machine translation (MT) is directly linked to its evaluation in order to both compare different MT system outputs and analyse system errors so that they can be addressed and corrected. As a consequence, MT evaluation has become increasingly important and popular in the last decade, leading to the development of MT evaluation metrics aiming at automatically assessing MT output. Most of these metrics use reference translations in order to compare system output, and the most well-known and widely spread work at lexical level. In this study we describe and present a linguistically-motivated metric, VERTa, which aims at using and combining a wide variety of linguistic features at lexical, morphological, syntactic and semantic level. Before designing and developing VERTa a qualitative linguistic analysis of data was performed so as to identify the linguistic phenomena that an MT metric must consider (Comelles et al. 2017). In the present study we introduce VERTa's design and architecture and we report the experiments performed in order to develop the metric and to check the suitability and interaction of the linguistic information used. The experiments carried out go beyond traditional correlation scores and step towards a more qualitative approach based on linguistic analysis. Finally, in order to check the validity of the metric, an evaluation has been conducted comparing the metric's performance to that of other well-known state-of-theart MT metrics.

Language Resources and Evaluation, 2010

We describe a dataset containing 16,000 translations produced by four machine translation systems and manually annotated for quality by professional translators. This dataset can be used in a range of tasks assessing machine translation evaluation metrics, from basic correlation analysis to training and test of machine learning-based metrics. By providing a standard dataset for such tasks, we hope to encourage the development of better MT evaluation metrics.

In the conventional evaluation metrics of machine translation, considering less information about the translations usually makes the result not reasonable and low correlation with human judgments. On the other hand, using many external linguistic resources and tools (e.g. Part-of-speech tagging, morpheme, stemming, and synonyms) makes the metrics complicated, time-consuming and not universal due to that different languages have the different linguistic features. This paper proposes a novel evaluation metric employing rich and augmented factors without relying on any additional resource or tool. Experiments show that this novel metric yields the state-of-the-art correlation with human judgments compared with classic metrics BLEU, TER, Meteor-1.3 and two latest metrics (AMBER and MP4IBM1), which proves it a robust one by employing a feature-rich and model-independent approach.

In the conventional evaluation metrics of machine translation, considering less information about the translations usually makes the result not reasonable and low correlation with human judgments. On the other hand, using many external linguistic resources and tools (e.g. Part-ofspeech tagging, morpheme, stemming, and synonyms) makes the metrics complicated, timeconsuming and not universal due to that different languages have the different linguistic features. This paper proposes a novel evaluation metric employing rich and augmented factors without relying on any additional resource or tool. Experiments show that this novel metric yields the state-of-the-art correlation with human judgments compared with classic metrics BLEU, TER, Meteor-1.3 and two latest metrics (AMBER and MP4IBM1), which proves it a robust one by employing a feature-rich and model-independent approach. 1

We introduce BLANC, a family of dynamic, trainable evaluation metrics for machine translation. Flexible, parametrized models can be learned from past data and automatically optimized to correlate well with human judgments for different criteria (e.g. adequacy, fluency) using different correlation measures. Towards this end, we discuss ACS (all common skipngrams), a practical algorithm with trainable parameters that estimates referencecandidate translation overlap by computing a weighted sum of all common skipngrams in polynomial time. We show that the BLEU and ROUGE metric families are special cases of BLANC, and we compare correlations with human judgments across these three metric families. We analyze the algorithmic complexity of ACS and argue that it is more powerful in modeling both local meaning and sentence-level structure, while offering the same practicality as the established algorithms it generalizes.

… of the conference on Human Language …, 2005

We introduce BLANC, a family of dynamic, trainable evaluation metrics for machine translation. Flexible, parametrized models can be learned from past data and automatically optimized to correlate well with human judgments for different criteria (e.g. adequacy, fluency) using different correlation measures. Towards this end, we discuss ACS (all common skipngrams), a practical algorithm with trainable parameters that estimates referencecandidate translation overlap by computing a weighted sum of all common skipngrams in polynomial time. We show that the BLEU and ROUGE metric families are special cases of BLANC, and we compare correlations with human judgments across these three metric families. We analyze the algorithmic complexity of ACS and argue that it is more powerful in modeling both local meaning and sentence-level structure, while offering the same practicality as the established algorithms it generalizes.

arXiv: Computation and Language, 2016

We introduce the Machine Translation (MT) evaluation survey that contains both manual and automatic evaluation methods. The traditional human evaluation criteria mainly include the intelligibility, fidelity, fluency, adequacy, comprehension, and informativeness. The advanced human assessments include task-oriented measures, post-editing, segment ranking, and extended criteriea, etc. We classify the automatic evaluation methods into two categories, including lexical similarity scenario and linguistic features application. The lexical similarity methods contain edit distance, precision, recall, F-measure, and word order. The linguistic features can be divided into syntactic features and semantic features respectively. The syntactic features include part of speech tag, phrase types and sentence structures, and the semantic features include named entity, synonyms, textual entailment, paraphrase, semantic roles, and language models. The deep learning models for evaluation are very newly ...

Seventh Conference on Machine Translation, 2022

We employ a linguistically motivated challenge set in order to evaluate the state-of-theart machine translation metrics submitted to the Metrics Shared Task of the 7th Conference for Machine Translation. The challenge set includes about 20,000 items extracted from 145 MT systems for two language directions (German⇔English), covering more than 100 linguistically-motivated phenomena organized in 14 categories. The best performing metrics are YiSi-1, BERTScore and COMET-22 for German-English, and UniTE, UniTE-ref, MetricX-XL-DA19 and MetricX-XXL-DA19 for English-German. Metrics in both directions are performing worst when it comes to namedentities & terminology and particularly measuring units. Particularly in German-English they are weak at detecting issues at punctuation, polar questions, relative clauses, dates and idioms. In English-German, they perform worst at present progressive of transitive verbs, future II progressive of intransitive verbs, simple present perfect of ditransitive verbs and focus particles.