Papers by Masayuki Suzuki
Robust ASR Error Correction with Conservative Data Filtering (2024.emnlp-industry)
Copied to clipboard
| Challenge: | Error correction (EC) based on large language models is an emerging technology to enhance the performance of automatic speech recognition systems. |
| Approach: | They propose to pair large set of ASR hypotheses with gold references to improve linguistic acceptability over sources and be inferable from available context. |
| Outcome: | The proposed approach significantly reduces overcorrection and improves quality in out-of-domain (OOD) settings. |
All-words Word Sense Disambiguation Using Concept Embeddings (L18-1)
Copied to clipboard
| Challenge: | Existing work on all-words word sense disambiguation (all-word WSD) uses word embeddings to identify the senses of words in documents. |
| Approach: | They propose a new concept embedding method to predict target word senses . concept embeds are constructed from concept tag sequences created from previous predictions . |
| Outcome: | The proposed concept embeddings improve Japanese all-words word sense disambiguation task. |