Sentence-level Aggregation of Lexical Metrics Correlates Stronger with Human Judgements than Corpus-level Aggregation
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913662144348160 |
|---|---|
| author | Cavalin, Paulo Domingues, Pedro Henrique Pinhanez, Claudio |
| author_facet | Cavalin, Paulo Domingues, Pedro Henrique Pinhanez, Claudio |
| contents | In this paper we show that corpus-level aggregation hinders considerably the capability of lexical metrics to accurately evaluate machine translation (MT) systems. With empirical experiments we demonstrate that averaging individual segment-level scores can make metrics such as BLEU and chrF correlate much stronger with human judgements and make them behave considerably more similar to neural metrics such as COMET and BLEURT. We show that this difference exists because corpus- and segment-level aggregation differs considerably owing to the classical average of ratio versus ratio of averages Mathematical problem. Moreover, as we also show, such difference affects considerably the statistical robustness of corpus-level aggregation. Considering that neural metrics currently only cover a small set of sufficiently-resourced languages, the results in this paper can help make the evaluation of MT systems for low-resource languages more trustworthy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_12832 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Sentence-level Aggregation of Lexical Metrics Correlates Stronger with Human Judgements than Corpus-level Aggregation Cavalin, Paulo Domingues, Pedro Henrique Pinhanez, Claudio Computation and Language In this paper we show that corpus-level aggregation hinders considerably the capability of lexical metrics to accurately evaluate machine translation (MT) systems. With empirical experiments we demonstrate that averaging individual segment-level scores can make metrics such as BLEU and chrF correlate much stronger with human judgements and make them behave considerably more similar to neural metrics such as COMET and BLEURT. We show that this difference exists because corpus- and segment-level aggregation differs considerably owing to the classical average of ratio versus ratio of averages Mathematical problem. Moreover, as we also show, such difference affects considerably the statistical robustness of corpus-level aggregation. Considering that neural metrics currently only cover a small set of sufficiently-resourced languages, the results in this paper can help make the evaluation of MT systems for low-resource languages more trustworthy. |
| title | Sentence-level Aggregation of Lexical Metrics Correlates Stronger with Human Judgements than Corpus-level Aggregation |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2407.12832 |