Constrained Generation

Translate Meanings, Not Just Words: IdiomKB's Role in Optimizing Idiomatic Translation with Language Models

We propose a multilingual idiom KB (IdiomKB) developed using LLMs to facilitate better idiomatic translation by smaller models by retrieving idioms’ figurative meanings.

Shuang Li, Jiangjie Chen, Siyu Yuan, Xinyi Wu, Hao Yang, Shimin Tao, Yanghua Xiao

Translate Meanings, Not Just Words: IdiomKB's Role in Optimizing Idiomatic Translation with Language Models

Distilling Script Knowledge from Large Language Models for Constrained Language Planning

We propose an over-generate-then-filter approach to improve large language models (LLMs) on constrained language planning, and use it to distill a novel constrained language planning dataset, CoScript.

Siyu Yuan, Jiangjie Chen, Ziquan Fu, Xuyang Ge, Soham Shah, Charles Robert Jankowski, Yanghua Xiao, Deqing Yang

Distilling Script Knowledge from Large Language Models for Constrained Language Planning

Say What You Mean! Large Language Models Speak Too Positively about Negative Commonsense Knowledge

We find that large language models (LLMs) speak too positively about negative commonsense knowledge, which is caused by statistical shortcuts and negation reporting bias from language modeling pre-training.

Jiangjie Chen, Wei Shi, Ziquan Fu, Sijie Cheng, Lei Li, Yanghua Xiao

Say What You Mean! Large Language Models Speak Too Positively about Negative Commonsense Knowledge

Converge to the Truth: Factual Error Correction via Iterative Constrained Editing

Fact verification guided iterative factual error correction without the supervision from corretion.

Jiangjie Chen, Rui Xu, Wenxuan Zeng, Changzhi Sun, Lei Li, Yanghua Xiao

Converge to the Truth: Factual Error Correction via Iterative Constrained Editing

Unsupervised Explanation Generation via Correct Instantiations

Generate explanations for why a statement is wrong by prompting LLMs with the correct version of it.

Sijie Cheng, Zhiyong Wu, Jiangjie Chen, Zhixing Li, Yang Liu, Lingpeng Kong

Unsupervised Explanation Generation via Correct Instantiations

Neighbors Are Not Strangers: Improving Non-Autoregressive Translation under Low-Frequency Lexical Constraints

We present ACT, an algorithm that enhances non-autoregressive translation (NAT) by improving constraint preservation and translation quality, especially for rare constraints.

Chun Zeng, Jiangjie Chen, Tianyi Zhuang, Rui Xu, Hao Yang, Ying Qin, Shimin Tao, Yanghua Xiao

Neighbors Are Not Strangers: Improving Non-Autoregressive Translation under Low-Frequency Lexical Constraints

Unsupervised Editing for Counterfactual Stories

Creating counterfactual story endings (what-if stories) with unsupervised editing.

Jiangjie Chen, Chun Gan, Sijie Cheng, Hao Zhou, Yanghua Xiao, Lei Li

Unsupervised Editing for Counterfactual Stories

Ensuring Readability and Data-fidelity using Head-modifier Templates in Deep Type Description Generation

Building head-modifier templates to constrain entity type description generation.

Jiangjie Chen, Ao Wang, Haiyun Jiang, Suo Feng, Chenguang Li, Yanghua Xiao

Ensuring Readability and Data-fidelity using Head-modifier Templates in Deep Type Description Generation