Identifying Weaknesses in Machine Translation Metrics Through Minimum Bayes Risk Decoding: A Case Study for COMET.
Chantal AmrheinRico SennrichPublished in: AACL/IJCNLP (1) (2022)
Keyphrases
- machine translation
- bayes risk
- finite state transducers
- upper and lower bounds
- loss function
- information extraction
- natural language processing
- cross lingual
- cross language information retrieval
- target language
- reproducing kernel hilbert space
- machine translation system
- natural language
- distortion measure
- posterior probability
- regularization parameter
- statistical machine translation
- precision and recall
- support vector
- vector quantization
- image quality
- quality metrics
- semi supervised