Publications

2024

1. arXiv

   NegotiationToM: A Benchmark for Stress-testing Machine Theory of Mind on Negotiation Surrounding

   Chunkit Chan, Cheng Jiayang, Yauwai Yim, Zheye Deng, Wei Fan, Haoran Li, Xin Liu, Hongming Zhang, Weiqi Wang, and Yangqiu Song

   arXiv preprint arXiv:2404.13627, 2024

   Bib

   @article{chan2024negotiationtom,
     title = {NegotiationToM: A Benchmark for Stress-testing Machine Theory of Mind on Negotiation Surrounding},
     author = {Chan, Chunkit and Jiayang, Cheng and Yim, Yauwai and Deng, Zheye and Fan, Wei and Li, Haoran and Liu, Xin and Zhang, Hongming and Wang, Weiqi and Song, Yangqiu},
     journal = {arXiv preprint arXiv:2404.13627},
     year = {2024}
   }

2. arXiv

   Text-Tuple-Table: Towards Information Integration in Text-to-Table Generation via Global Tuple Extraction

   Zheye Deng, Chunkit Chan, Weiqi Wang, Yuxi Sun, Wei Fan, Tianshi Zheng, Yauwai Yim, and Yangqiu Song

   arXiv preprint arXiv:2404.14215, 2024

   Bib

   @article{deng2024text,
     title = {Text-Tuple-Table: Towards Information Integration in Text-to-Table Generation via Global Tuple Extraction},
     author = {Deng, Zheye and Chan, Chunkit and Wang, Weiqi and Sun, Yuxi and Fan, Wei and Zheng, Tianshi and Yim, Yauwai and Song, Yangqiu},
     journal = {arXiv preprint arXiv:2404.14215},
     year = {2024}
   }

3. Arxiv

   CANDLE: Iterative Conceptualization and Instantiation Distillation from Large Language Models for Commonsense Reasoning

   Weiqi Wang, Tianqing Fang, Chunyang Li, Haochen Shi, Wenxuan Ding, Baixuan Xu, Zhaowei Wang, Jiaxin Bai, Xin Liu, Jiayang Cheng, Chunkit Chan, and Yangqiu Song

   2024

   Bib

   @misc{wang2024candle,
     title = {CANDLE: Iterative Conceptualization and Instantiation Distillation from Large Language Models for Commonsense Reasoning},
     author = {Wang, Weiqi and Fang, Tianqing and Li, Chunyang and Shi, Haochen and Ding, Wenxuan and Xu, Baixuan and Wang, Zhaowei and Bai, Jiaxin and Liu, Xin and Cheng, Jiayang and Chan, Chunkit and Song, Yangqiu},
     year = {2024},
     eprint = {2401.07286},
     archiveprefix = {arXiv},
     primaryclass = {cs.CL}
   }

4. Main-COLING

   EventGround: Narrative Reasoning by Grounding to Eventuality-centric Knowledge Graphs

   Jiayang Cheng, Lin Qiu, Chunkit Chan, Xin Liu, Yangqiu Song, and Zheng Zhang

   In , 2024

   Bib

   @inproceedings{Cheng2024EventGroundNR,
     title = {EventGround: Narrative Reasoning by Grounding to Eventuality-centric Knowledge Graphs},
     author = {Cheng, Jiayang and Qiu, Lin and Chan, Chunkit and Liu, Xin and Song, Yangqiu and Zhang, Zheng},
     year = {2024},
     url = {https://api.semanticscholar.org/CorpusID:268819426}
   }

2023

1. Findings-EACL

   ChatGPT Evaluation on Sentence Level Relations: A Focus on Temporal, Causal, and Discourse Relations

   Chunkit Chan, Jiayang Cheng, Weiqi Wang, Yuxin Jiang, Tianqing Fang, Xin Liu, and Yangqiu Song

   Findings-EACL2024, 2023

   Bib

   @article{DBLP:journals/corr/abs-2304-14827,
     author = {Chan, Chunkit and Cheng, Jiayang and Wang, Weiqi and Jiang, Yuxin and Fang, Tianqing and Liu, Xin and Song, Yangqiu},
     title = {ChatGPT Evaluation on Sentence Level Relations: {A} Focus on Temporal,
                       Causal, and Discourse Relations},
     journal = {Findings-EACL2024},
     volume = {abs/2304.14827},
     year = {2023},
     url = {https://doi.org/10.48550/arXiv.2304.14827},
     doi = {10.48550/arXiv.2304.14827},
     eprinttype = {arXiv},
     eprint = {2304.14827},
     timestamp = {Thu, 04 May 2023 15:47:42 +0200},
     biburl = {https://dblp.org/rec/journals/corr/abs-2304-14827.bib},
     bibsource = {dblp computer science bibliography, https://dblp.org}
   }

2. Findings-ACL

   DiscoPrompt: Path Prediction Prompt Tuning for Implicit Discourse Relation Recognition

   Chunkit Chan, Xin Liu, Jiayang Cheng, Zihan Li, Yangqiu Song, Ginny Y. Wong, and Simon See

   In Findings of the Association for Computational Linguistics: ACL 2023, Toronto, Canada, July 9-14, 2023, 2023

   Bib

   @inproceedings{DBLP:conf/acl/ChanLCLSWS23,
     author = {Chan, Chunkit and Liu, Xin and Cheng, Jiayang and Li, Zihan and Song, Yangqiu and Wong, Ginny Y. and See, Simon},
     editor = {Rogers, Anna and Boyd{-}Graber, Jordan L. and Okazaki, Naoaki},
     title = {DiscoPrompt: Path Prediction Prompt Tuning for Implicit Discourse
                       Relation Recognition},
     booktitle = {Findings of the Association for Computational Linguistics: {ACL} 2023,
                       Toronto, Canada, July 9-14, 2023},
     pages = {35--57},
     publisher = {Association for Computational Linguistics},
     year = {2023},
     url = {https://doi.org/10.18653/v1/2023.findings-acl.4},
     doi = {10.18653/v1/2023.findings-acl.4},
     timestamp = {Thu, 10 Aug 2023 12:35:51 +0200},
     biburl = {https://dblp.org/rec/conf/acl/ChanLCLSWS23.bib},
     bibsource = {dblp computer science bibliography, https://dblp.org}
   }

3. Oral-AACL

   Self-Consistent Narrative Prompts on Abductive Natural Language Inference

   Chunkit Chan, Xin Liu, Tsz Ho Chan, Jiayang Cheng, Yangqiu Song, Ginny Wong, and Simon See

   In Proceedings of the 13th International Joint Conference on Natural Language Processing and the 3rd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers), Nov 2023

   Bib

   @inproceedings{chan-etal-2023-self,
     title = {Self-Consistent Narrative Prompts on Abductive Natural Language Inference},
     author = {Chan, Chunkit and Liu, Xin and Chan, Tsz Ho and Cheng, Jiayang and Song, Yangqiu and Wong, Ginny and See, Simon},
     editor = {Park, Jong C. and Arase, Yuki and Hu, Baotian and Lu, Wei and Wijaya, Derry and Purwarianti, Ayu and Krisnadhi, Adila Alfa},
     booktitle = {Proceedings of the 13th International Joint Conference on Natural Language Processing and the 3rd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)},
     month = nov,
     year = {2023},
     address = {Nusa Dua, Bali},
     publisher = {Association for Computational Linguistics},
     url = {https://aclanthology.org/2023.ijcnlp-main.67},
     pages = {1040--1057}
   }

4. Oral-EMNLP

   Lion: Adversarial Distillation of Proprietary Large Language Models

   Yuxin Jiang, Chunkit Chan, Mingyang Chen, and Wei Wang

   In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, Dec 2023

   Abs Bib

   The practice of transferring knowledge from a sophisticated, proprietary large language model (LLM) to a compact, open-source LLM has garnered considerable attention. Previous works have focused on a unidirectional knowledge distillation way by aligning the responses of the student model with those of the teacher model to a set of instructions. Nevertheless, they overlooked the possibility of incorporating any “feedback”–identifying challenging instructions where the student model’s performance falls short–to boost the student model’s proficiency iteratively. To this end, we propose a novel adversarial distillation framework for a more efficient knowledge transfer. Leveraging the versatile role adaptability of LLMs, we prompt the teacher model to identify “hard” instructions and generate new “hard” instructions for the student model, creating a three-stage adversarial loop of imitation, discrimination, and generation. By applying this adversarial framework, we successfully transfer knowledge from ChatGPT to a student model (named Lion), using a mere 70k training data. Our results show that Lion-13B not only achieves comparable open-ended generation capabilities to ChatGPT but surpasses conventional state-of-the-art (SOTA) instruction-tuned models like Vicuna-13B by 55.4% in challenging zero-shot reasoning benchmarks such as BIG-Bench Hard (BBH) and 16.7% on AGIEval.

   @inproceedings{jiang-etal-2023-lion,
     title = {Lion: Adversarial Distillation of Proprietary Large Language Models},
     author = {Jiang, Yuxin and Chan, Chunkit and Chen, Mingyang and Wang, Wei},
     editor = {Bouamor, Houda and Pino, Juan and Bali, Kalika},
     booktitle = {Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing},
     month = dec,
     year = {2023},
     address = {Singapore},
     publisher = {Association for Computational Linguistics},
     url = {https://aclanthology.org/2023.emnlp-main.189},
     doi = {10.18653/v1/2023.emnlp-main.189},
     pages = {3134--3154}
   }

5. Main-EMNLP

   StoryAnalogy: Deriving Story-level Analogies from Large Language Models to Unlock Analogical Understanding

   Cheng Jiayang, Lin Qiu, Tsz Chan, Tianqing Fang, Weiqi Wang, Chunkit Chan, Dongyu Ru, Qipeng Guo, Hongming Zhang, Yangqiu Song, Yue Zhang, and Zheng Zhang

   In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, Dec 2023

   Abs Bib

   Analogy-making between narratives is crucial for human reasoning. In this paper, we evaluate the ability to identify and generate analogies by constructing a first-of-its-kind large-scale story-level analogy corpus, StoryAnalogy, which contains 24K story pairs from diverse domains with human annotations on two similarities from the extended Structure-Mapping Theory. We design a set of tests on StoryAnalogy, presenting the first evaluation of story-level analogy identification and generation. Interestingly, we find that the analogy identification tasks are incredibly difficult not only for sentence embedding models but also for the recent large language models (LLMs) such as ChatGPT and LLaMa. ChatGPT, for example, only achieved around 30% accuracy in multiple-choice questions (compared to over 85% accuracy for humans). Furthermore, we observe that the data in StoryAnalogy can improve the quality of analogy generation in LLMs, where a fine-tuned FlanT5-xxl model achieves comparable performance to zero-shot ChatGPT.

   @inproceedings{jiayang-etal-2023-storyanalogy,
     title = {{S}tory{A}nalogy: Deriving Story-level Analogies from Large Language Models to Unlock Analogical Understanding},
     author = {Jiayang, Cheng and Qiu, Lin and Chan, Tsz and Fang, Tianqing and Wang, Weiqi and Chan, Chunkit and Ru, Dongyu and Guo, Qipeng and Zhang, Hongming and Song, Yangqiu and Zhang, Yue and Zhang, Zheng},
     editor = {Bouamor, Houda and Pino, Juan and Bali, Kalika},
     booktitle = {Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing},
     month = dec,
     year = {2023},
     address = {Singapore},
     publisher = {Association for Computational Linguistics},
     url = {https://aclanthology.org/2023.emnlp-main.706},
     doi = {10.18653/v1/2023.emnlp-main.706},
     pages = {11518--11537}
   }

6. Arxiv

   P-Bench: A Multi-level Privacy Evaluation Benchmark for Language Models

   Haoran Li, Dadi Guo, Donghao Li, Wei Fan, Qi Hu, Xin Liu, Chunkit Chan, Duanyi Yao, and Yangqiu Song

   Dec 2023

   Bib

   @article{li2023pbench,
     title = {P-Bench: A Multi-level Privacy Evaluation Benchmark for Language Models},
     author = {Li, Haoran and Guo, Dadi and Li, Donghao and Fan, Wei and Hu, Qi and Liu, Xin and Chan, Chunkit and Yao, Duanyi and Song, Yangqiu},
     year = {2023},
     eprint = {2311.04044},
     archiveprefix = {arXiv},
     primaryclass = {cs.CL}
   }

7. Arxiv

   Privacy in Large Language Models: Attacks, Defenses and Future Directions

   Haoran Li, Yulin Chen, Jinglong Luo, Yan Kang, Xiaojin Zhang, Qi Hu, Chunkit Chan, and Yangqiu Song

   Dec 2023

   Bib

   @article{li2023privacy,
     title = {Privacy in Large Language Models: Attacks, Defenses and Future Directions},
     author = {Li, Haoran and Chen, Yulin and Luo, Jinglong and Kang, Yan and Zhang, Xiaojin and Hu, Qi and Chan, Chunkit and Song, Yangqiu},
     year = {2023},
     eprint = {2310.10383},
     archiveprefix = {arXiv},
     primaryclass = {cs.CL}
   }