HalOmi: A Manually Annotated Benchmark for Multilingual Hallucination and Omission Detection in Machine Translation.
David DaleElena VoitaJanice LamPrangthip HansantiChristophe RopersElahe KalbassiCynthia GaoLoïc BarraultMarta R. Costa-jussàPublished in: EMNLP (2023)
Keyphrases
- machine translation
- manually annotated
- cross lingual
- language independent
- cross language information retrieval
- language resources
- relation extraction
- chinese english
- multilingual documents
- information extraction
- ground truth
- machine translation system
- natural language processing
- target language
- parallel corpus
- natural language
- statistical machine translation
- word alignment
- query translation
- parallel corpora
- named entities
- information retrieval
- knowledge base
- cross language
- digital libraries
- hidden markov models
- semantic features
- named entity recognition
- wordnet
- n gram