Papers by Dong Nguyen
On learning and representing social meaning in NLP: a sociolinguistic perspective (2021.naacl-main)
Copied to clipboard
| Challenge: | linguistic variation allows for the expression of social meaning, information about the social background and identity of the language user. |
| Approach: | They introduce the concept of social meaning to NLP and discuss how sociolinguistics can inform work on representation learning in NLP. |
| Outcome: | The proposed model can be used to learn social meaning in NLP and identify key challenges. |
tBERT: Topic Models and BERT Joining Forces for Semantic Similarity Detection (2020.acl-main)
Copied to clipboard
| Challenge: | Recent pretrained contextual representations such as ELMo and BERT have led to impressive performance gains across a variety of NLP tasks, including semantic similarity detection. |
| Approach: | They propose a topic-informed BERT-based architecture for pairwise semantic similarity detection that adds topic information to pretrained contextual representations such as BERT. |
| Outcome: | The proposed model outperforms existing models on a variety of English language datasets and is highly performant. |
Gradient-Boosted Decision Tree for Listwise Context Model in Multimodal Review Helpfulness Prediction (2023.findings-acl)
Copied to clipboard
| Challenge: | Existing studies have shown that FCNNs perform inefficient splitting for review features, making it difficult to clearly differentiate helpful from unhelpful reviews. |
| Approach: | They propose a listwise attention network that captures the MRHP ranking context and a pairwise optimization objective that enhances model generalization. |
| Outcome: | The proposed framework achieves state-of-the-art results and polished generalization performance on two large-scale MRHP benchmark datasets. |
Measuring the Instability of Fine-Tuning (2023.acl-long)
Copied to clipboard
| Challenge: | Many previous studies have investigated fine-tuning pre-trained language models on downstream tasks with varying random seeds, but they only used the standard deviation of performance scores (SD) as their measure, which is a narrow characterization of instability. |
| Approach: | They propose a systematic evaluation framework for the standard deviation of performance scores (SD) and six other measures quantifying instability of different granularity levels. |
| Outcome: | The proposed framework will be used to evaluate the validity of these measures and to improve them. |
Modeling Dynamic Topics in Chain-Free Fashion by Evolution-Tracking Contrastive Learning and Unassociated Word Exclusion (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing dynamic topic models lack the ability to reveal the evolution of topics . Existing models suffer from repetitive topic and unassociated topic issues . |
| Approach: | They propose a new evolution-tracking contrastive learning method that builds the similarity relations among dynamic topics and an unassociated word exclusion method to avoid unassociated topics. |
| Outcome: | The proposed model outperforms state-of-the-art models on downstream tasks and is robust to evolution intensities. |
HateCheck: Functional Tests for Hate Speech Detection Models (2021.acl-long)
Copied to clipboard
| Challenge: | Hate speech detection models are evaluated by measuring their performance on held-out test data using metrics such as accuracy and F1 score. |
| Approach: | They propose a suite of functional tests for hate speech detection models that measure model performance on held-out test data and then craft test cases to validate their quality. |
| Outcome: | The proposed tests show that the proposed models perform poorly on a small set of widely-used hate speech datasets. |
PSET: a Phonetics-Semantics Evaluation Testbed (2025.emnlp-main)
Copied to clipboard
| Challenge: | Phonetic embeddings are a powerful tool for capturing the meaning of text . they are not ideal for tasks centered on sound, such as finding sound analogies . |
| Approach: | They propose a phonetic-semantics evaluation testbed to evaluate phonetic embeddings . they use phonetic embedded models to test phonetic models . |
| Outcome: | The phonetic embeddings solve the task with varying degrees of success . the phonetic-based embeddables perform better than the other models . |
Crossing Linguistic Horizons: Finetuning and Comprehensive Evaluation of Vietnamese Large Language Models (2024.findings-naacl)
Copied to clipboard
| Challenge: | Existing open-source LLMs exhibit limited effectiveness in processing Vietnamese . lack of systematic benchmark datasets and metrics tailored for Vietnamese LLM evaluation exacerbates these issues. |
| Approach: | They propose to fine tune LLMs specifically for Vietnamese and develop a framework for evaluation . they find that larger models introduce more biases and uncalibrated outputs . |
| Outcome: | The proposed framework finetunes LLMs specifically for Vietnamese and provides a framework for evaluation . |
Template-based Abstractive Microblog Opinion Summarization (2022.tacl-1)
Copied to clipboard
| Challenge: | Existing work on Twitter uses extractive summarization to filter through information, but this approach often includes incomplete or redundant information. |
| Approach: | They propose to use Twitter data to generate 3100 gold-standard opinion summaries. |
| Outcome: | The proposed method outperforms previous work on extractive summarization models and fine-tunes to improve performance. |
Room to Glo: A Systematic Comparison of Semantic Change Detection Approaches with Word Embeddings (D19-1)
Copied to clipboard
| Challenge: | Word embeddings are increasingly used for automatic detection of semantic change, but a robust evaluation and systematic comparison of the choices involved has been lacking. |
| Approach: | They propose a new evaluation framework for semantic change detection using whole time series and a Twitter dataset spanning 5.5 years. |
| Outcome: | The proposed framework shows that using whole time series is preferable over continuously trained embeddings for long time periods and that the reference point matters. |
Disentangling the Roles of Representation and Selection in Data Pruning (2025.acl-long)
Copied to clipboard
| Challenge: | Existing methods for data pruning involve many different design choices, which have not been systematically studied. |
| Approach: | They decompose data pruning into two key components: data representation and selection algorithm . theoretical and empirical results highlight crucial role of representations . |
| Outcome: | The proposed method can be used to train models with less data. |
We Need to Measure Data Diversity in NLP — Better and Broader (2025.emnlp-main)
Copied to clipboard
| Challenge: | Language models exhibit remarkable natural language understanding and generation capabilities, but they have serious flaws, such as societal biases and spurious correlations. |
| Approach: | They argue that interdisciplinary perspectives are essential for developing more fine-grained and valid measures of data diversity. |
| Outcome: | The proposed measures are based on interdisciplinary perspectives and include a variety of datasets. |
Does It Capture STEL? A Modular, Similarity-based Linguistic Style Evaluation Framework (2021.emnlp-main)
Copied to clipboard
| Challenge: | linguistic style is an integral part of natural language, but evaluation methods for style measures are rare, often task-specific and usually do not control for content. |
| Approach: | They propose a modular, fine-grained and content-controlled similarity-based STyle EvaLuation framework to test the performance of any model that can compare two sentences on style. |
| Outcome: | The proposed model outperforms simple versions of commonly used style measures like 3-grams, punctuation frequency and LIWC-based approaches. |
Automated Generation of Accurate & Fluent Medical X-ray Reports (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing medical report generation efforts focus on producing human-readable reports, yet the generated text may not be well aligned to the clinical facts. |
| Approach: | They propose to automate the generation of medical reports from chest X-ray image inputs . medical reports are the primary medium, which physicians communicate findings from scans - authors say . |
| Outcome: | The proposed method achieves fluency and clinical accuracy on common metrics. |
Do Word Embeddings Capture Spelling Variation? (2020.coling-main)
Copied to clipboard
| Challenge: | Using word embeddings, we analyze spelling variation in word embeds trained on Twitter and Reddit data. |
| Approach: | They propose a new perspective on the analysis of word embeddings by focusing on spelling variation. |
| Outcome: | The proposed analysis shows that word embeddings encode spelling variation patterns of various types to some extent, even when trained using the skipgram model. |
BERT, are you paying attention? Attention regularization with human-annotated rationales (2026.eacl-long)
Copied to clipboard
| Challenge: | Attention regularisation aims to supervise the attention patterns in language models like BERT. |
| Approach: | They compare regularisation on human rationales with random tokens to find that human-annotated rationale is better at reducing model sensitivity to spurious correlations. |
| Outcome: | The proposed regularisation method improves model performance and model robustness, but not with human-annotated rationales. |
DemaFormer: Damped Exponential Moving Average Transformer with Energy-Based Modeling for Temporal Language Grounding (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Temporal Language Grounding (TLG) is a task to determine temporal boundaries of video moments that correspond to a language query. |
| Approach: | They propose an energy-based model framework to explicitly learn moment-query distributions. |
| Outcome: | The proposed model outperforms the state-of-the-art models on four public temporal language grounding datasets. |
Aiming beyond the Obvious: Identifying Non-Obvious Cases in Semantic Similarity Datasets (P19-1)
Copied to clipboard
| Challenge: | Existing datasets for scoring text pairs in terms of semantic similarity contain instances whose resolution differs according to the degree of difficulty. |
| Approach: | They propose to use lexical overlap to distinguish obvious from non-obvious text pairs by focusing on item difficulty and ground-truth labels to characterise existing datasets. |
| Outcome: | The proposed models are based on lexical overlap and ground-truth labels and focus on cases of similarity which require more complex inference. |
Comparing Automatic and Human Evaluation of Local Explanations for Text Classification (N18-1)
Copied to clipboard
| Challenge: | Text classification models are becoming increasingly complex and opaque, however for many applications it is essential that the models are interpretable. |
| Approach: | They propose to use automatic word deletion to generate local explanations for a text classification model by crowdsourcing the evaluation using a crowdsourced experiment. |
| Outcome: | The proposed evaluations of local explanations using crowdsourcing and automatic measures correlate with the results. |
Tokenization is Sensitive to Language Variation (2025.findings-acl)
Copied to clipboard
| Challenge: | Variation in language is often linked to regional, social, and contextual factors. |
| Approach: | They propose a method to estimate tokenizer impact on downstream LLM performance . they pre-train BERT models with the popular Byte-Pair Encoding algorithm . |
| Outcome: | The proposed model improves on Rényi efficiency and other metrics on language variation. |
Assessing the Reliability of Word Embedding Gender Bias Measures (2021.emnlp-main)
Copied to clipboard
| Challenge: | Various measures have been proposed to quantify human-like social biases in word embeddings, but they can suffer from measurement error. |
| Approach: | They propose to assess the reliability of word embedding gender bias measures by examining their reliability across different choices of random seeds, scoring rules and words. |
| Outcome: | The proposed measures can suffer from measurement error, and the results inform better design of word embedding gender bias measures. |
Introducing CAD: the Contextual Abuse Dataset (2021.naacl-main)
Copied to clipboard
| Challenge: | Detecting and classifying online abuse is a complex and nuanced task, despite many advances in the power and availability of computational tools. |
| Approach: | They propose to annotate a reddit conversation thread with six distinct primary and secondary categories and an expert-driven group-adjudication process for high quality annotations. |
| Outcome: | The proposed dataset contains six distinct primary and secondary categories and uses an expert-driven group-adjudication process for high quality annotations. |
FTFT: Efficient and Robust Fine-Tuning by Transferring Training Dynamics (2025.coling-main)
Copied to clipboard
| Challenge: | Despite the success of fine-tuning Pre-trained Language Models, they remain susceptible to out-of-distribution input. |
| Approach: | They propose a novel approach that fine-tunes Pre-trained Language Models by transFerring Training dynamics (FTFT) FTFT uses more efficient reference models and aggressive early stopping . |
| Outcome: | The proposed approach improves the robustness of fine-tuned PLMs while reducing training costs. |
What’s Mine becomes Yours: Defining, Annotating and Detecting Context-Dependent Paraphrases in News Interview Dialogs (2024.emnlp-main)
Copied to clipboard
| Challenge: | a dataset of utterance pairs from NPR and CNN is used to classify paraphrases in dialog. |
| Approach: | They propose a dataset annotated for context-dependent paraphrases and develop a training for crowd-workers to classify paraphrase in dialog. |
| Outcome: | The proposed dataset contains 5,581 annotations on 600 utterance pairs. |