Posts by Collection

portfolio

publications

Lying Through One’s Teeth: A Study on Verbal Leakage Cues

Published in EMNLP, 2021

We study verbal leakage cues to understand the effect of the data construction method on their significance, and examine the relationship between such cues and models' validity. Result shows that data with audio statements and lie-based annotations indicate a greater number of strong verbal leakage cue categories, and shows that models trained on a dataset with more strong verbal leakage cue categories yield superior results.

Recommended citation: Min-Hsuan Yeh and Lun-Wei Ku. (2021). "Lying Through One’s Teeth: A Study on Verbal Leakage Cues," to appear in Proceedings of the The 2021 Conference on Empirical Methods in Natural Language Processing (EMNLP). https://aclanthology.org/2021.emnlp-main.370.pdf

Multi-VQG: Generating Engaging Questions for Multiple Images

Published in EMNLP, 2022

We propose a new task, Multi-VQG, which aims to generate engaging questions for multiple images. We introduce a new dataset, MVQG, which contains arround 30,000 question and image sequence pairs. We also propose both end-to-end and dual-staged models extended from VL-T5 to generate questions with story information. We evaluate our models on MVQG and show that models with explicit story information yield better results.

Recommended citation: Min-Hsuan Yeh, Vicent Chen, Ting-Hao 'Kenneth' Haung, and Lun-Wei Ku. (2022). "Multi-VQG: Generating Engaging Questions for Multiple Images," to appear in Proceedings of the The 2022 Conference on Empirical Methods in Natural Language Processing (EMNLP). https://arxiv.org/abs/2211.07441

talks

teaching

Teaching experience 1

Undergraduate course, University 1, Department, 2014

This is a description of a teaching experience. You can use markdown like any other post.

Teaching experience 2

Workshop, University 1, Department, 2015

This is a description of a teaching experience. You can use markdown like any other post.