Papers by Thang Vu
Teaching a Multilingual Large Language Model to Understand Multilingual Speech via Multi-Instructional Training (2024.findings-naacl)
Copied to clipboard
| Challenge: | Recent advances in language modeling have led to the emergence of large language models capable ofvarious natural language processing tasks. |
| Approach: | They propose a multi-instructional training approach that integrates a large language model with a speech encoder to harness the capabilities of LLMs for speech recognition and beyond. |
| Outcome: | The proposed model can be trained and aligned with a multilingual LLM on 1900 hours of transcribed data from 139 languages. |
AmericasNLI: Evaluating Zero-shot Natural Language Understanding of Pretrained Multilingual Models in Truly Low-resource Languages (2022.acl-long)
Copied to clipboard
Abteen Ebrahimi, Manuel Mager, Arturo Oncevay, Vishrav Chaudhary, Luis Chiruzzo, Angela Fan, John Ortega, Ricardo Ramos, Annette Rios, Ivan Vladimir Meza Ruiz, Gustavo Giménez-Lugo, Elisabeth Mager, Graham Neubig, Alexis Palmer, Rolando Coto-Solano, Thang Vu, Katharina Kann
| Challenge: | Pretrained multilingual models can perform cross-lingual transfer in a zero-shot setting, even for unseen languages. |
| Approach: | They propose to extend XNLI to 10 indigenous languages of the Americas and test multiple zero-shot and translation-based approaches. |
| Outcome: | The proposed model can perform cross-lingual transfer in a zero-shot setting even for languages unseen during pretraining. |
FreshLLMs: Refreshing Large Language Models with Search Engine Augmentation (2024.findings-acl)
Copied to clipboard
Tu Vu, Mohit Iyyer, Xuezhi Wang, Noah Constant, Jerry Wei, Jason Wei, Chris Tar, Yun-Hsuan Sung, Denny Zhou, Quoc Le, Thang Luong
| Challenge: | Modern large language models often "hallucinate" plausible but factually incorrect information, which reduces their trustworthiness especially in settings where accurate and up-to-date information is critical. |
| Approach: | They develop a human evaluation procedure to measure correctness and hallucination and use it to benchmark both closed and open-source LLMs. |
| Outcome: | The proposed method outperforms both competing search engine-augmented prompting methods and commercial systems on search-augmented QA. |
Understanding the Role of Mental Models in User Interaction with an Adaptive Dialog Agent (2025.findings-naacl)
Copied to clipboard
| Challenge: | Adaptive dialog systems can help users align their behavior with user expectations, but there is little research into what mental models users form when interacting with a task-oriented dialog system. |
| Approach: | They propose to use a publicly available dataset to explore user mental models of dialog systems to better align with users' mental models. |
| Outcome: | The proposed model can improve dialog efficiency, success, and user perception of the interaction, even when done implicitly. |
Language-Agnostic Meta-Learning for Low-Resource Text-to-Speech with Articulatory Features (2022.acl-long)
Copied to clipboard
| Challenge: | Recent advances in text-to-speech systems allow for speech synthesis with unprecedented quality and controllability. |
| Approach: | They use embeddings derived from articulatory vectors rather than phoneme identities to learn phoneme representations that hold across languages. |
| Outcome: | The proposed models fine-tuned on 30 minutes of data in a previously unseen language with language agnostic meta learning. |
Data Augmentation Techniques for Machine Translation of Code-Switched Texts: A Comparative Study (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Code-switching (CSW) text generation is a popular solution to address data scarcity. |
| Approach: | They compare linguistic theories, lexical replacements and back-translation approaches to Egyptian Arabic-English CSW. |
| Outcome: | The proposed methods perform best on machine translation and quality evaluation. |
BPE vs. Morphological Segmentation: A Case Study on Machine Translation of Four Polysynthetic Languages (2022.findings-acl)
Copied to clipboard
| Challenge: | Morphologically rich polysynthetic languages present a challenge for NLP systems due to data sparsity. |
| Approach: | They propose to use subword segmentation to reduce data sparsity in polysynthetic languages . they compare supervised and unsupervised morphological segmentation methods to Byte-Pair Encodings . |
| Outcome: | The proposed methods outperform BPEs in MT tasks for all language pairs except for Nahuatl . the proposed methods are more efficient than supervised methods, but less sparse in fusional languages. |
Meta Learning for Natural Language Processing: A Survey (2022.naacl-main)
Copied to clipboard
| Challenge: | Meta-learning is an emerging field in machine learning, but there is no systematic survey of these approaches in NLP. |
| Approach: | They propose to introduce meta-learning and the common approaches and summarize their work and review their work in the NLP community. |
| Outcome: | The proposed methods improve performance in many NLP tasks but are limited to domains, languages, countries, or styles. |
PhotoshopQuiA: A Corpus of Non-Factoid Questions and Answers for Why-Question Answering (L18-1)
Copied to clipboard
| Challenge: | Community Question Answering web sites are used for non-factoid question answering . however, there is a scarcity of available datasets for this task . cnn.com's john m. sutter is releasing a dataset for why-QA . |
| Approach: | They propose a dataset of 2,854 why-question and answer(s) pairs related to Adobe Photoshop usage from five CQA web sites. |
| Outcome: | The new dataset is the first English dataset for Why-QA that focuses on a product . it can be used to build Why-Q systems, evaluate approaches and develop new models . |
Please note that I’m just an AI: Analysis of Behavior Patterns of LLMs in (Non-)offensive Speech Identification (2024.emnlp-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) are becoming a part of our everyday lives by being used as tools for information search, content creation, writing assistance, and many more. |
| Approach: | They propose to use Large Language Models to detect offensive online language in applications with social risk, such as late-life companions and online content moderators. |
| Outcome: | The proposed models fail to detect offensive language and are therefore unsuitable for use in social applications such as late-life companions and online content moderators. |