Papers by Thuy-Trang Vu
Direct Evaluation of Chain-of-Thought in Multi-hop Reasoning with Knowledge Graphs (2024.findings-acl)
Copied to clipboard
| Challenge: | Prior research on evaluating large language models focused on answer accuracy, neglecting the correctness of the generated CoT. |
| Approach: | They propose a discriminative and generative CoT evaluation paradigm to assess LLMs’ knowledge of reasoning and the accuracy of the generated CoT. |
| Outcome: | The proposed evaluation paradigm assesses LLMs’ knowledge of reasoning and the accuracy of the generated CoT. |
Extending LLMs to New Languages: A Case Study of Llama and Persian Adaptation (2025.coling-main)
Copied to clipboard
| Challenge: | Large language models (LLMs) are mainly trained on English data and struggle with low-resource languages. |
| Approach: | They propose to add a new language to Llama to improve classification accuracy for Persian tasks by aligning representations through bilingual pretraining and instruction datasets. |
| Outcome: | The proposed model performs on generation and classification tasks with no adverse impact and sometimes even improvements on English tasks. |
Automatic Post-Editing of Machine Translation: A Neural Programmer-Interpreter Approach (D18-1)
Copied to clipboard
| Challenge: | Existing approaches to inducing APE have suffered from over-correction, where the APE system tends to keep the machine translated text without any modification. |
| Approach: | They propose a neural programmer-interpreter approach to automated post-editing (APE) that mimics human perform post- editing using discrete edit operations . their model outperforms previous neural models for inducing PE programs on the WMT17 APE task for German-English up to +1 BLEU score and -0.7 TER scores. |
| Outcome: | The proposed model outperforms previous neural models for inducing PE programs on the WMT17 APE task for German-English up to +1 BLEU score and -0.7 TER scores. |
Discrete Minds in a Continuous World: Do Language Models Know Time Passes? (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Large Language Models excel at temporal reasoning tasks, but their ability to perceive the passage of time remains unexplored. |
| Approach: | They propose a Token-Time Hypothesis to test whether LLMs perceive the passage of time . they also propose an interactive navigation challenge to examine how LLM responds to time pressure . |
| Outcome: | The proposed model can map discrete token counts to wall-clock time and validate this through a dialogue duration judgment task. |
SCAR: Data Selection via Style Consistency-Aware Response Ranking for Efficient Instruction-Tuning of Large Language Models (2025.acl-long)
Copied to clipboard
| Challenge: | Recent studies show that manually ensuring a consistent response style and maintaining high data quality can significantly improve the performance of fine-tuned Large Language Models (LLMs). |
| Approach: | They introduce a style-aware response ranking system that prioritizes instruction-response pairs based on their stylistic consistency. |
| Outcome: | The proposed model matches or surpasses models trained on the entire dataset in coding and open-ended question-answering benchmarks. |
MixLoRA-DSI: Dynamically Expandable Mixture-of-LoRA Experts for Rehearsal-Free Generative Retrieval over Dynamic Corpora (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing approaches to update model-based indexes with new documents are expensive and require expensive retraining. |
| Approach: | They propose a framework that combines an expandable mixture of Low-Rank Adaptation experts with a layer-wise out-of-distribution-driven expansion strategy. |
| Outcome: | Experiments on NQ320k and MS MARCO Passage show that the proposed framework outperforms full-model update baselines with minimal parameter overhead and substantially lower training costs. |
CONGRAD: Conflicting Gradient Filtering for Multilingual Preference Alignment (2026.eacl-long)
Copied to clipboard
Jiangnan Li, Thuy-Trang Vu, Christian Herold, Amirhossein Tebbifakhr, Shahram Khadivi, Gholamreza Haffari
| Challenge: | Naive joint training of large language models can suffer from negative interference. |
| Approach: | They propose a filtering method that aggregates cross-lingually beneficial gradients and filters for those with high cross-linguistic affinity. |
| Outcome: | The proposed method outperforms baselines in both seen and unseen languages with minimal alignment tax. |
LiveCultureBench: a Multi-Agent, Multi-Cultural Benchmark for Large Language Models in Dynamic Social Simulations (2026.acl-long)
Copied to clipboard
| Challenge: | Large language models (LLMs) are increasingly deployed as autonomous agents . evaluations focus primarily on task success rather than cultural appropriateness or reliability. |
| Approach: | They propose a multi-cultural, dynamic benchmark that embeds large language models as agents in a simulated town and evaluates them on task completion and adherence to socio-cultural norms. |
| Outcome: | The proposed model evaluates LLMs on task completion and adherence to socio-cultural norms across models and cultural profiles. |
Generalised Unsupervised Domain Adaptation of Neural Machine Translation with Cross-Lingual Data Selection (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing work on unsupervised domain adaptation of neural machine translation assumes access to monolingual text in either the source or target language in the new domain. |
| Approach: | They propose a method to extract in-domain sentences from a large generic monolingual corpus from 'missing' text. |
| Outcome: | The proposed method outperforms baselines up to +1.5 BLEU score on five diverse domains in three language pairs and a real-world translation scenario. |
Domain Generalisation of NMT: Fusing Adapters with Leave-One-Domain-Out Training (2022.findings-acl)
Copied to clipboard
| Challenge: | Recent advances in neural machine translation (NMT) research have found NMT models sensitive to distribution shift and adversarial examples. |
| Approach: | They propose a leave-one-domain-out training strategy that learns to combine domain-specific parameters to avoid information leaking. |
| Outcome: | The proposed method outperforms baselines on three language pairs on average. |
Koala: An Index for Quantifying Overlaps with Pre-training Corpora (2023.emnlp-demo)
Copied to clipboard
| Challenge: | Recent studies have shown that large language models can be influenced by the frequency of overlap between pre-training corpora. |
| Approach: | They propose to search over large pre-training corpora using lossless compressed suffix arrays with highly efficient compression rate and search support. |
| Outcome: | Koala is a searchable index over large pre-training corpora using lossless compressed suffix arrays with highly efficient compression rate and search support. |
Mixture-of-Skills: Learning to Optimize Data Usage for Fine-Tuning Large Language Models (2024.emnlp-main)
Copied to clipboard
| Challenge: | Large language models are fine-tuned on diverse datasets to develop a range of skills . each skill has unique characteristics, and datasets are heterogeneous and imbalanced . a general, model-agnostic, reinforcement learning framework is proposed to optimize data usage . |
| Approach: | They propose a general, model-agnostic, reinforcement learning framework that optimizes data usage automatically during the fine-tuning process. |
| Outcome: | The proposed framework optimizes data usage automatically during the fine-tuning process. |
Discourse Graph Guided Document Translation with Large Language Models (2026.eacl-long)
Copied to clipboard
| Challenge: | Recent agentic machine translation systems mitigate context window constraints but require substantial computational resources and are sensitive to memory retrieval strategies. |
| Approach: | They propose a framework that explicitly models inter-chunk relationships through structured discourse graphs and selectively conditions each translation segment on relevant graph neighbourhoods rather than sequential or exhaustive context. |
| Outcome: | The proposed framework surpasses strong baselines in translation quality and terminology consistency while incurring significantly lower token overhead. |
Learning How to Active Learn by Dreaming (P19-1)
Copied to clipboard
| Challenge: | Recent active learning methods are limited when the data distribution of learning problems vary. |
| Approach: | They propose a wake-and-dream-based active learning method that learns the AL policy directly on the target domain of interest by using wake and dream cycles. |
| Outcome: | The proposed method improves on cross-domain and cross-lingual tasks. |
Proverbs Run in Pairs: Evaluating Proverb Translation Capability of Large Language Model (2025.findings-acl)
Copied to clipboard
| Challenge: | Recent research has demonstrated that large language models (LLMs) can translate cultural elements in languages such as idioms and proverbs. |
| Approach: | They propose to use large language models to translate culturally rooted proverbs in conversation and between languages with similar cultural backgrounds to compare their results. |
| Outcome: | The proposed models can achieve good translation between languages with similar cultural backgrounds and outperform NMT models in proverb translation. |
Exploring the Potential of Multimodal LLM with Knowledge-Intensive Multimodal ASR (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Recent advances in multimodal large language models have made significant progress in integrating information across various modalities, yet real-world applications in educational and scientific domains remain challenging. |
| Approach: | They propose a task that focuses on transcribing scientific conference videos by leveraging visual information from slides to enhance the accuracy of technical terminologies. |
| Outcome: | The proposed framework improves transcript quality through post-editing and improves performance over speech-only baselines. |
Towards Inference-time Scaling for Continuous Space Reasoning (2026.findings-acl)
Copied to clipboard
| Challenge: | Recent advances in reasoning large language models have expanded along training and inferencetime dimensions. |
| Approach: | They propose to use COCONUT (CITATION) continuous space reasoning LM as the backbone to generate diverse reasoning paths through dropout-based sampling. |
| Outcome: | The proposed method could enable performance gains similar to those observed in the discrete space, but only marginally improves in the continuous space. |
Effective Unsupervised Domain Adaptation with Adversarially Trained Language Models (2020.emnlp-main)
Copied to clipboard
| Challenge: | Recent work has shown the importance of training contextualised word embedding models on the domain of the target task of interest. |
| Approach: | They propose a masking strategy which adversarially masks out those tokens which are harder to reconstruct by the underlying MLM. |
| Outcome: | The proposed training strategy outperforms random masking on six unsupervised domain adaptation tasks and achieves up to +1.64 F1 score improvements. |