Merge pull request #2 from WING-NUS/hugoblox-import-publications

Hugo Blox Builder - Import latest publications

content/publication/aksu-etal-2021-velocidapter/cite.bib

@@ -0,0 +1,26 @@
+@inproceedings{aksu-etal-2021-velocidapter,
+ abstract = {We introduce a synthetic dialogue generation framework, Velocidapter, which addresses the corpus availability problem for dialogue comprehension. Velocidapter augments datasets by simulating synthetic conversations for a task-oriented dialogue domain, requiring a small amount of bootstrapping work for each new domain. We evaluate the efficacy of our framework on a task-oriented dialogue comprehension dataset, MRCWOZ, which we curate by annotating questions for slots in the restaurant, taxi, and hotel domains of the MultiWOZ 2.2 dataset (Zang et al., 2020). We run experiments within a low-resource setting, where we pretrain a model on SQuAD, fine-tuning it on either a small original data or on the synthetic data generated by our framework. Velocidapter shows significant improvements using both the transformer-based BERTBase and BiDAF as base models. We further show that the framework is easy to use by novice users and conclude that Velocidapter can greatly help training over task-oriented dialogues, especially for low-resourced emerging domains.},
+ address = {Singapore and Online},
+ author = {Aksu, Ibrahim Taha  and
+Liu, Zhengyuan  and
+Kan, Min-Yen  and
+Chen, Nancy},
+ booktitle = {Proceedings of the 22nd Annual Meeting of the Special Interest Group on Discourse and Dialogue},
+ doi = {10.18653/v1/2021.sigdial-1.14},
+ editor = {Li, Haizhou  and
+Levow, Gina-Anne  and
+Yu, Zhou  and
+Gupta, Chitralekha  and
+Sisman, Berrak  and
+Cai, Siqi  and
+Vandyke, David  and
+Dethlefs, Nina  and
+Wu, Yan  and
+Li, Junyi Jessy},
+ month = {July},
+ pages = {133--143},
+ publisher = {Association for Computational Linguistics},
+ title = {Velocidapter: Task-oriented Dialogue Comprehension Modeling Pairing Synthetic Text Generation with Domain Adaptation},
+ url = {https://aclanthology.org/2021.sigdial-1.14},
+ year = {2021}
+}

content/publication/aksu-etal-2021-velocidapter/index.md

@@ -0,0 +1,32 @@
+---
+title: 'Velocidapter: Task-oriented Dialogue Comprehension Modeling Pairing Synthetic
+  Text Generation with Domain Adaptation'
+authors:
+- Ibrahim Taha Aksu
+- Zhengyuan Liu
+- Min-Yen Kan
+- Nancy Chen
+date: '2021-07-01'
+publishDate: '2024-07-05T17:09:42.645613Z'
+publication_types:
+- paper-conference
+publication: '*Proceedings of the 22nd Annual Meeting of the Special Interest Group
+  on Discourse and Dialogue*'
+doi: 10.18653/v1/2021.sigdial-1.14
+abstract: We introduce a synthetic dialogue generation framework, Velocidapter, which
+  addresses the corpus availability problem for dialogue comprehension. Velocidapter
+  augments datasets by simulating synthetic conversations for a task-oriented dialogue
+  domain, requiring a small amount of bootstrapping work for each new domain. We evaluate
+  the efficacy of our framework on a task-oriented dialogue comprehension dataset,
+  MRCWOZ, which we curate by annotating questions for slots in the restaurant, taxi,
+  and hotel domains of the MultiWOZ 2.2 dataset (Zang et al., 2020). We run experiments
+  within a low-resource setting, where we pretrain a model on SQuAD, fine-tuning it
+  on either a small original data or on the synthetic data generated by our framework.
+  Velocidapter shows significant improvements using both the transformer-based BERTBase
+  and BiDAF as base models. We further show that the framework is easy to use by novice
+  users and conclude that Velocidapter can greatly help training over task-oriented
+  dialogues, especially for low-resourced emerging domains.
+links:
+- name: URL
+  url: https://aclanthology.org/2021.sigdial-1.14
+---

content/publication/aksu-etal-2022-n/cite.bib

@@ -0,0 +1,19 @@
+@inproceedings{aksu-etal-2022-n,
+ abstract = {Augmentation of task-oriented dialogues has followed standard methods used for plain-text such as back-translation, word-level manipulation, and paraphrasing despite its richly annotated structure. In this work, we introduce an augmentation framework that utilizes belief state annotations to match turns from various dialogues and form new synthetic dialogues in a bottom-up manner. Unlike other augmentation strategies, it operates with as few as five examples. Our augmentation strategy yields significant improvements when both adapting a DST model to a new domain, and when adapting a language model to the DST task, on evaluations with TRADE and TOD-BERT models. Further analysis shows that our model performs better on seen values during training, and it is also more robust to unseen values. We conclude that exploiting belief state annotations enhances dialogue augmentation and results in improved models in n-shot training scenarios.},
+ address = {Dublin, Ireland},
+ author = {Aksu, Ibrahim  and
+Liu, Zhengyuan  and
+Kan, Min-Yen  and
+Chen, Nancy},
+ booktitle = {Findings of the Association for Computational Linguistics: ACL 2022},
+ doi = {10.18653/v1/2022.findings-acl.131},
+ editor = {Muresan, Smaranda  and
+Nakov, Preslav  and
+Villavicencio, Aline},
+ month = {May},
+ pages = {1659--1671},
+ publisher = {Association for Computational Linguistics},
+ title = {N-Shot Learning for Augmenting Task-Oriented Dialogue State Tracking},
+ url = {https://aclanthology.org/2022.findings-acl.131},
+ year = {2022}
+}

content/publication/aksu-etal-2022-n/index.md

@@ -0,0 +1,29 @@
+---
+title: N-Shot Learning for Augmenting Task-Oriented Dialogue State Tracking
+authors:
+- Ibrahim Aksu
+- Zhengyuan Liu
+- Min-Yen Kan
+- Nancy Chen
+date: '2022-05-01'
+publishDate: '2024-07-05T17:09:42.588862Z'
+publication_types:
+- paper-conference
+publication: '*Findings of the Association for Computational Linguistics: ACL 2022*'
+doi: 10.18653/v1/2022.findings-acl.131
+abstract: Augmentation of task-oriented dialogues has followed standard methods used
+  for plain-text such as back-translation, word-level manipulation, and paraphrasing
+  despite its richly annotated structure. In this work, we introduce an augmentation
+  framework that utilizes belief state annotations to match turns from various dialogues
+  and form new synthetic dialogues in a bottom-up manner. Unlike other augmentation
+  strategies, it operates with as few as five examples. Our augmentation strategy
+  yields significant improvements when both adapting a DST model to a new domain,
+  and when adapting a language model to the DST task, on evaluations with TRADE and
+  TOD-BERT models. Further analysis shows that our model performs better on seen values
+  during training, and it is also more robust to unseen values. We conclude that exploiting
+  belief state annotations enhances dialogue augmentation and results in improved
+  models in n-shot training scenarios.
+links:
+- name: URL
+  url: https://aclanthology.org/2022.findings-acl.131
+---

content/publication/dou-etal-2022-towards/cite.bib

@@ -0,0 +1,24 @@
+@inproceedings{dou-etal-2022-towards,
+ abstract = {In this paper, we study the problem of knowledge-intensive text-to-SQL, in which domain knowledge is necessary to parse expert questions into SQL queries over domain-specific tables. We formalize this scenario by building a new benchmark KnowSQL consisting of domain-specific questions covering various domains. We then address this problem by representing formulaic knowledge rather than by annotating additional data examples. More concretely, we construct a formulaic knowledge bank as a domain knowledge base and propose a framework (ReGrouP) to leverage this formulaic knowledge during parsing. Experiments using ReGrouP demonstrate a significant 28.2% improvement overall on KnowSQL.},
+ address = {Abu Dhabi, United Arab Emirates},
+ author = {Dou, Longxu  and
+Gao, Yan  and
+Liu, Xuqi  and
+Pan, Mingyang  and
+Wang, Dingzirui  and
+Che, Wanxiang  and
+Zhan, Dechen  and
+Kan, Min-Yen  and
+Lou, Jian-Guang},
+ booktitle = {Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing},
+ doi = {10.18653/v1/2022.emnlp-main.350},
+ editor = {Goldberg, Yoav  and
+Kozareva, Zornitsa  and
+Zhang, Yue},
+ month = {December},
+ pages = {5240--5253},
+ publisher = {Association for Computational Linguistics},
+ title = {Towards Knowledge-Intensive Text-to-SQL Semantic Parsing with Formulaic Knowledge},
+ url = {https://aclanthology.org/2022.emnlp-main.350},
+ year = {2022}
+}

content/publication/dou-etal-2022-towards/index.md

@@ -0,0 +1,32 @@
+---
+title: Towards Knowledge-Intensive Text-to-SQL Semantic Parsing with Formulaic Knowledge
+authors:
+- Longxu Dou
+- Yan Gao
+- Xuqi Liu
+- Mingyang Pan
+- Dingzirui Wang
+- Wanxiang Che
+- Dechen Zhan
+- Min-Yen Kan
+- Jian-Guang Lou
+date: '2022-12-01'
+publishDate: '2024-07-05T17:09:42.603419Z'
+publication_types:
+- paper-conference
+publication: '*Proceedings of the 2022 Conference on Empirical Methods in Natural
+  Language Processing*'
+doi: 10.18653/v1/2022.emnlp-main.350
+abstract: In this paper, we study the problem of knowledge-intensive text-to-SQL,
+  in which domain knowledge is necessary to parse expert questions into SQL queries
+  over domain-specific tables. We formalize this scenario by building a new benchmark
+  KnowSQL consisting of domain-specific questions covering various domains. We then
+  address this problem by representing formulaic knowledge rather than by annotating
+  additional data examples. More concretely, we construct a formulaic knowledge bank
+  as a domain knowledge base and propose a framework (ReGrouP) to leverage this formulaic
+  knowledge during parsing. Experiments using ReGrouP demonstrate a significant 28.2%
+  improvement overall on KnowSQL.
+links:
+- name: URL
+  url: https://aclanthology.org/2022.emnlp-main.350
+---

content/publication/han-etal-2022-mm/cite.bib

@@ -0,0 +1,19 @@
+@inproceedings{han-etal-2022-mm,
+ abstract = {Existing multimodal tasks mostly target at the complete input modality setting, i.e., each modality is either complete or completely missing in both training and test sets. However, the randomly missing situations have still been underexplored. In this paper, we present a novel approach named MM-Align to address the missing-modality inference problem. Concretely, we propose 1) an alignment dynamics learning module based on the theory of optimal transport (OT) for missing data imputation; 2) a denoising training algorithm to enhance the quality of imputation as well as the accuracy of model predictions. Compared with previous generative methods which devote to restoring the missing inputs, MM-Align learns to capture and imitate the alignment dynamics between modality sequences. Results of comprehensive experiments on two multimodal tasks empirically demonstrate that our method can perform more accurate and faster inference and alleviate the overfitting issue under different missing conditions.},
+ address = {Abu Dhabi, United Arab Emirates},
+ author = {Han, Wei  and
+Chen, Hui  and
+Kan, Min-Yen  and
+Poria, Soujanya},
+ booktitle = {Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing},
+ doi = {10.18653/v1/2022.emnlp-main.717},
+ editor = {Goldberg, Yoav  and
+Kozareva, Zornitsa  and
+Zhang, Yue},
+ month = {December},
+ pages = {10498--10511},
+ publisher = {Association for Computational Linguistics},
+ title = {MM-Align: Learning Optimal Transport-based Alignment Dynamics for Fast and Accurate Inference on Missing Modality Sequences},
+ url = {https://aclanthology.org/2022.emnlp-main.717},
+ year = {2022}
+}

content/publication/han-etal-2022-mm/index.md

@@ -0,0 +1,32 @@
+---
+title: 'MM-Align: Learning Optimal Transport-based Alignment Dynamics for Fast and
+  Accurate Inference on Missing Modality Sequences'
+authors:
+- Wei Han
+- Hui Chen
+- Min-Yen Kan
+- Soujanya Poria
+date: '2022-12-01'
+publishDate: '2024-07-05T17:09:42.610472Z'
+publication_types:
+- paper-conference
+publication: '*Proceedings of the 2022 Conference on Empirical Methods in Natural
+  Language Processing*'
+doi: 10.18653/v1/2022.emnlp-main.717
+abstract: Existing multimodal tasks mostly target at the complete input modality setting,
+  i.e., each modality is either complete or completely missing in both training and
+  test sets. However, the randomly missing situations have still been underexplored.
+  In this paper, we present a novel approach named MM-Align to address the missing-modality
+  inference problem. Concretely, we propose 1) an alignment dynamics learning module
+  based on the theory of optimal transport (OT) for missing data imputation; 2) a
+  denoising training algorithm to enhance the quality of imputation as well as the
+  accuracy of model predictions. Compared with previous generative methods which devote
+  to restoring the missing inputs, MM-Align learns to capture and imitate the alignment
+  dynamics between modality sequences. Results of comprehensive experiments on two
+  multimodal tasks empirically demonstrate that our method can perform more accurate
+  and faster inference and alleviate the overfitting issue under different missing
+  conditions.
+links:
+- name: URL
+  url: https://aclanthology.org/2022.emnlp-main.717
+---

content/publication/han-etal-2024-self/cite.bib

@@ -0,0 +1,18 @@
+@inproceedings{han-etal-2024-self,
+ abstract = {Image--text models (ITMs) is the prevalent architecture to solve video question--answering tasks, which requires only a few input frames to save huge computational cost compared to video--language models.However, we find existent ITM video question--answering solutions either 1) adopt simplistic and unintentional sampling strategies, which may miss key frames to offer the answer clues; or 2) sample a large number of frames into divided groups, which the computational sources can not accommodate. In this work, we aim at an efficient sampling method towards the few-frame situations.We first summarize a family of prior sampling methods based on question--frame correlation into a unified one, dubbed *Most Implied Frames* (MIF). Through some primary results and analysis, Through analysis, we form a hypothesis that question-aware sampling is not necessary, from which we further propose the other method *Most Dominant Frames* (MDF).Experimental results on four public datasets and three advanced ITMs demonstrate that our proposed strategies can boost the performance for image--text pretrained models, and have a wide application scenario in terms of model architectures and dataset types. Our code is available at https://github.com/declare-lab/Sealingr̆lhttps://github.com/declare-lab/Sealing.},
+ address = {Mexico City, Mexico},
+ author = {Han, Wei  and
+Chen, Hui  and
+Kan, Min-Yen  and
+Poria, Soujanya},
+ booktitle = {Findings of the Association for Computational Linguistics: NAACL 2024},
+ editor = {Duh, Kevin  and
+Gomez, Helena  and
+Bethard, Steven},
+ month = {June},
+ pages = {2522--2534},
+ publisher = {Association for Computational Linguistics},
+ title = {Self-Adaptive Sampling for Accurate Video Question Answering on Image Text Models},
+ url = {https://aclanthology.org/2024.findings-naacl.162},
+ year = {2024}
+}

content/publication/han-etal-2024-self/index.md

@@ -0,0 +1,32 @@
+---
+title: Self-Adaptive Sampling for Accurate Video Question Answering on Image Text
+  Models
+authors:
+- Wei Han
+- Hui Chen
+- Min-Yen Kan
+- Soujanya Poria
+date: '2024-06-01'
+publishDate: '2024-07-05T17:09:42.578623Z'
+publication_types:
+- paper-conference
+publication: '*Findings of the Association for Computational Linguistics: NAACL 2024*'
+abstract: Image--text models (ITMs) is the prevalent architecture to solve video question--answering
+  tasks, which requires only a few input frames to save huge computational cost compared
+  to video--language models.However, we find existent ITM video question--answering
+  solutions either 1) adopt simplistic and unintentional sampling strategies, which
+  may miss key frames to offer the answer clues; or 2) sample a large number of frames
+  into divided groups, which the computational sources can not accommodate. In this
+  work, we aim at an efficient sampling method towards the few-frame situations.We
+  first summarize a family of prior sampling methods based on question--frame correlation
+  into a unified one, dubbed *Most Implied Frames* (MIF). Through some primary results
+  and analysis, Through analysis, we form a hypothesis that question-aware sampling
+  is not necessary, from which we further propose the other method *Most Dominant
+  Frames* (MDF).Experimental results on four public datasets and three advanced ITMs
+  demonstrate that our proposed strategies can boost the performance for image--text
+  pretrained models, and have a wide application scenario in terms of model architectures
+  and dataset types. Our code is available at https://github.com/declare-lab/Sealingr̆lhttps://github.com/declare-lab/Sealing.
+links:
+- name: URL
+  url: https://aclanthology.org/2024.findings-naacl.162
+---

content/publication/jain-etal-2022-comparative/cite.bib

@@ -0,0 +1,21 @@
+@inproceedings{jain-etal-2022-comparative,
+ abstract = {We model products′ reviews to generate comparative responses consisting of positive and negative experiences regarding the product. Specifically, we generate a single-sentence, comparative response from a given positive and a negative opinion. We contribute the first dataset for this task of Comparative Snippet Generation from contrasting opinions regarding a product, and an analysis of performance of a pre-trained BERT model to generate such snippets.},
+ address = {Dublin, Ireland},
+ author = {Jain, Saurabh  and
+Miao, Yisong  and
+Kan, Min-Yen},
+ booktitle = {Proceedings of the Fifth Workshop on e-Commerce and NLP (ECNLP 5)},
+ doi = {10.18653/v1/2022.ecnlp-1.7},
+ editor = {Malmasi, Shervin  and
+Rokhlenko, Oleg  and
+Ueffing, Nicola  and
+Guy, Ido  and
+Agichtein, Eugene  and
+Kallumadi, Surya},
+ month = {May},
+ pages = {49--57},
+ publisher = {Association for Computational Linguistics},
+ title = {Comparative Snippet Generation},
+ url = {https://aclanthology.org/2022.ecnlp-1.7},
+ year = {2022}
+}

content/publication/jain-etal-2022-comparative/index.md

@@ -0,0 +1,22 @@
+---
+title: Comparative Snippet Generation
+authors:
+- Saurabh Jain
+- Yisong Miao
+- Min-Yen Kan
+date: '2022-05-01'
+publishDate: '2024-07-05T17:09:42.617512Z'
+publication_types:
+- paper-conference
+publication: '*Proceedings of the Fifth Workshop on e-Commerce and NLP (ECNLP 5)*'
+doi: 10.18653/v1/2022.ecnlp-1.7
+abstract: We model products′ reviews to generate comparative responses consisting
+  of positive and negative experiences regarding the product. Specifically, we generate
+  a single-sentence, comparative response from a given positive and a negative opinion.
+  We contribute the first dataset for this task of Comparative Snippet Generation
+  from contrasting opinions regarding a product, and an analysis of performance of
+  a pre-trained BERT model to generate such snippets.
+links:
+- name: URL
+  url: https://aclanthology.org/2022.ecnlp-1.7
+---

content/publication/qin-etal-2022-gl/cite.bib

@@ -0,0 +1,22 @@
+@inproceedings{qin-etal-2022-gl,
+ abstract = {Due to high data demands of current methods, attention to zero-shot cross-lingual spoken language understanding (SLU) has grown, as such approaches greatly reduce human annotation effort. However, existing models solely rely on shared parameters, which can only perform implicit alignment across languages. We present Global-Local Contrastive Learning Framework (GL-CLeF) to address this shortcoming. Specifically, we employ contrastive learning, leveraging bilingual dictionaries to construct multilingual views of the same utterance, then encourage their representations to be more similar than negative example pairs, which achieves to explicitly align representations of similar sentences across languages. In addition, a key step in GL-CLeF is a proposed Local and Global component, which achieves a fine-grained cross-lingual transfer (i.e., sentence-level Local intent transfer, token-level Local slot transfer, and semantic-level Global transfer across intent and slot). Experiments on MultiATIS++ show that GL-CLeF achieves the best performance and successfully pulls representations of similar sentences across languages closer.},
+ address = {Dublin, Ireland},
+ author = {Qin, Libo  and
+Chen, Qiguang  and
+Xie, Tianbao  and
+Li, Qixin  and
+Lou, Jian-Guang  and
+Che, Wanxiang  and
+Kan, Min-Yen},
+ booktitle = {Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
+ doi = {10.18653/v1/2022.acl-long.191},
+ editor = {Muresan, Smaranda  and
+Nakov, Preslav  and
+Villavicencio, Aline},
+ month = {May},
+ pages = {2677--2686},
+ publisher = {Association for Computational Linguistics},
+ title = {GL-CLeF: A Global--Local Contrastive Learning Framework for Cross-lingual Spoken Language Understanding},
+ url = {https://aclanthology.org/2022.acl-long.191},
+ year = {2022}
+}