Papers by Guanyi Chen
DGST: a Dual-Generator Network for Text Style Transfer (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing studies on text style transfer focus on altering sentiment words to preserve attribute-independent information. |
| Approach: | They propose a Dual-Generator network architecture for text Style Transfer using two generators. |
| Outcome: | The proposed model performs better than existing models on Yelp and IMDb datasets. |
Computational Modelling of Plurality and Definiteness in Chinese Noun Phrases (2024.lrec-main)
Copied to clipboard
| Challenge: | linguists have suggested that some languages are "cooler" than others because of their contexts. |
| Approach: | They propose to omit plurality and definiteness markers in Chinese noun phrases . they build a corpus of Chinese NPs accompanied by its context . |
| Outcome: | The proposed model predicts the plurality and definiteness of Chinese noun phrases (NPs) it shows that speakers drop plurality markers very frequently, and that they are more likely to drop pronouns . |
On the Human-level Performance of Visual Question Answering (2025.coling-main)
Copied to clipboard
| Challenge: | Visual7W has been widely used in assessing multiple-choice visual question-answering systems. |
| Approach: | They replicated a human experiment on Visual7W to examine the human-level performance of VQA. |
| Outcome: | The results show that the better a model performs on Visual7W, the better it aligns with human-level intelligence. |
Highly Efficient Knowledge Graph Embedding Learning with Orthogonal Procrustes Analysis (2021.naacl-main)
Copied to clipboard
| Challenge: | Knowledge Graph Embeddings (KGEs) have been explored in recent years due to their promise for a wide range of applications. |
| Approach: | They propose a KGE framework which can reduce the training time and carbon footprint by orders of magnitudes compared with state-of-the-art approaches. |
| Outcome: | The proposed framework reduces the training time and carbon footprint by orders of magnitudes compared with state-of-the-art approaches while producing competitive performance. |
Improving Variational Autoencoder for Text Modelling with Timestep-Wise Regularisation (2020.coling-main)
Copied to clipboard
| Challenge: | Variational Autoencoders (VAEs) have been widely used in text modelling but posterior collapse is a problem when RNN-based models are employed. |
| Approach: | They propose a timestep-wise regularisation VAE architecture which can effectively avoid posterior collapse when used in text modelling. |
| Outcome: | The proposed model avoids posterior collapse and can be applied to any RNN-based VAE model. |
Non-neural Models Matter: a Re-evaluation of Neural Referring Expression Generation Systems (2022.acl-long)
Copied to clipboard
| Challenge: | In recent years, neural models have outperformed rule-based and classic approaches in NLG. |
| Approach: | They evaluate two English datasets and evaluate their performance using automatic and human evaluations. |
| Outcome: | The proposed model outperforms rule-based and classic approaches on two English datasets and is compared with human-based models. |
MMChat: Multi-Modal Chat Dataset on Social Media (2022.lrec-1)
Copied to clipboard
| Challenge: | Incorporating multi-modal contexts in conversation is important for developing engaging dialogue systems. |
| Approach: | They propose a large scale Chinese multi-modal dialogue corpus that contains image-grounded dialogues from real conversations on social media. |
| Outcome: | The proposed model can handle sparsity issues in dialogue generation tasks by incorporating image features. |
Understanding the Use of Quantifiers in Mandarin (2022.findings-aacl)
Copied to clipboard
| Challenge: | a corpus of short texts in Mandarin is analyzed to examine the "coolness" hypothesis . quantified expressions are used to describe short texts, but are not as informative as English . |
| Approach: | They propose a corpus of Mandarin in which quantified expressions figure prominently. |
| Outcome: | The proposed corpus of short texts in Mandarin is compared with an English corpus. |
GPTEval: A Survey on Assessments of ChatGPT and GPT-4 (2024.lrec-main)
Copied to clipboard
| Challenge: | emergence of ChatGPT has generated speculation about its potential to disrupt social and economic systems. |
| Approach: | They analyze prior assessments of ChatGPT and GPT-4 to analyze their language and reasoning abilities, scientific knowledge, ethical considerations and existing evaluation methods. |
| Outcome: | The proposed model performs satisfactorily in science knowledge and can answer open questions. |
Intrinsic Task-based Evaluation for Referring Expression Generation (2024.acl-long)
Copied to clipboard
| Challenge: | Referring Expression Generation (REG) models generate referring expressions that refer to referents at different points in a discourse. |
| Approach: | They propose to use a purely ratings-based human evaluation to evaluate REG models by completing two meta-level tasks. |
| Outcome: | The proposed evaluation makes the models more reliable and discriminable, and improves the quality of the REs. |