Papers by Ruitao Feng
Retrieval-Augmented Modular Prompt Tuning for Low-Resource Data-to-Text Generation (2024.lrec-main)
Copied to clipboard
| Challenge: | Data-to-text generation methods are often limited by data sparsity and lack of training data. |
| Approach: | They propose a retrieval-augmented modular prompt tuning method that generates texts with few hallucinations from structured data inputs. |
| Outcome: | The proposed method generates texts with few hallucinations and achieves state-of-the-art performance on a dataset for drone handover message generation. |
ART: The Alternating Reading Task Corpus for Speech Entrainment and Imitation (2024.lrec-main)
Copied to clipboard
Zheng Byron Yuan, Dorina de Jong, Ruitao Feng, Štefan Beňuš, Noël Nguyen, Róbert Sabo, Luciano Fadiga, Alessandro D’Ausilio
| Challenge: | The Alternating Reading Task (ART) Corpus is a collection of dyadic sentence readings for studying the entrainment and imitation behaviour in speech communication. |
| Approach: | They propose to use dyadic sentence reading to study entrainment and imitation in speech communication. |
| Outcome: | The proposed study includes three conditions and three subcorpora encompassing French-, Italian-, and Slovak-accented English. |