Jacky ChunKit Chan

PhD student @HKUST in NLP supervised by @yqsong. My research interests are Theory of Mind and Social Intelligence. @HKUSTKnowComp #NLProc

[Jacky ChunKit Chan @Jacky_chunkit]

We quantitatively evaluate the performance of ChatGPT on inter-sentential relations such as temporal relations, causal relations, and discourse relations. More details of ChatGPT performance can be found in our arxiv paper. https://t.co/WymjwGDtan

[Jacky ChunKit Chan @Jacky_chunkit]

Excit to share our two papers have been accepted to #EMNLP2023 main track. One is "🦁Lion: Adversarial Distillation of Proprietary Large Language Models". We propose an adversarial distillation framework for transferring knowledge from a proprietary LLM to a open source LLM.

[Yuxin Jiang @ ICLR2025 @Yuxin_Jiang_]

🎉Our paper has been accepted to Main Conference of #EMNLP2023: "🦁Lion: Adversarial Distillation of Proprietary Large Language Models". We propose an adversarial distillation framework for efficiently transferring knowledge from a proprietary LLM to a small, open-source LLM.

[Jacky ChunKit Chan @Jacky_chunkit]

Besides, we propose a general self-consistent framework that considers various narrative sequences (e.g., linear narrative and reverse chronology) for guiding the pre-trained language model in understanding the narrative context of input.
https://t.co/qnE1NE5KFt

[Jacky ChunKit Chan @Jacky_chunkit]

🚨Our new paper📷has been accepted to IJCNLP-AACL 2023: "Self-Consistent Narrative Prompts on Abductive Natural Language Inference". We propose a prompt tuning model which takes self-consistency and intersentential coherence into consideration for αNLI task.

[Jiaxin Bai @JiaxinBai2]

Excited to share our #KDD23 paper "Knowledge Graph Reasoning over Entities and Numerical Values" In this paper, we discuss the problem of how to answer complex logical queries over the knowledge graphs with numerical types. Welcome to discuss!

[Jacky ChunKit Chan @Jacky_chunkit]

Thanks to all my co-authors @yqsong @SeanLiu0x60 @jchengaj @lizihanov17 @HKUSTKnowComp
@NVIDIAAI for helping me with this work!
Check the paper below!
https://t.co/xtifL94NIH

[Jacky ChunKit Chan @Jacky_chunkit]

🚨Our new paper🚨has been accepted to Findings of #ACL2023: "DiscoPrompt: Path Prediction Prompt Tuning for Implicit Discourse Relation Recognition". We try to address how to inject labels’ hierarchical structure information into pre-trained language models via prompt tuning.

[Jacky ChunKit Chan @Jacky_chunkit]

We model the Implicit Discourse Relation Recognition (IDRR) problem as the path prediction problem that predicts the top-level relations, second-level types, and connectives at the same time.

[Jacky ChunKit Chan @Jacky_chunkit]

Thanks for @yqsong @SeanLiu0x60 @TFang229 @MightyWeaver2 and other co-authors to complete this work on #ChatGPT.

[Jacky ChunKit Chan @Jacky_chunkit]

We quantitatively evaluate the performance of ChatGPT on inter-sentential relations such as temporal relations, causal relations, and discourse relations. More details of ChatGPT performance can be found in our arxiv paper. https://t.co/WymjwGDtan