no code implementations • EMNLP 2018 • Fuli Luo, Tianyu Liu, Zexue He, Qiaolin Xia, Zhifang Sui, Baobao Chang
The goal of Word Sense Disambiguation (WSD) is to identify the correct meaning of a word in the particular context.
no code implementations • ICLR 2019 • Haohan Wang, Zexue He, Zachary C. Lipton, Eric P. Xing
We test our method on the battery of standard domain generalization data sets and, interestingly, achieve comparable or better performance as compared to other domain generalization methods that explicitly require samples from the target distribution for training.
Ranked #113 on Domain Generalization on PACS
1 code implementation • Findings (EMNLP) 2021 • Zexue He, Bodhisattwa Prasad Majumder, Julian McAuley
Written language carries explicit and implicit biases that can distract from meaningful signals.
1 code implementation • Findings (EMNLP) 2021 • An Yan, Zexue He, Xing Lu, Jiang Du, Eric Chang, Amilcare Gentili, Julian McAuley, Chun-Nan Hsu
Radiology report generation aims at generating descriptive text from radiology images automatically, which may present an opportunity to improve radiology reporting and interpretation.
no code implementations • 6 Mar 2022 • Canwen Xu, Zexue He, Zhankui He, Julian McAuley
Language models (LMs) can reproduce (or amplify) toxic language seen during training, which poses a risk to their practical application.
1 code implementation • 14 Oct 2022 • Zexue He, Yu Wang, Julian McAuley, Bodhisattwa Prasad Majumder
However, when sensitive information is semantically entangled with the task information of the input, e. g., gender information is predictive for a profession, a fair trade-off between task performance and bias mitigation is difficult to achieve.
no code implementations • 14 Oct 2022 • Bodhisattwa Prasad Majumder, Zexue He, Julian McAuley
In the other setup, human feedback was able to disentangle associated bias and predictive information from the input leading to superior bias mitigation and improved task performance (4-5%) simultaneously.
no code implementations • 19 Dec 2022 • Zexue He, Graeme Blackwood, Rameswar Panda, Julian McAuley, Rogerio Feris
Pre-training models with large crawled corpora can lead to issues such as toxicity and bias, as well as copyright and privacy concerns.
no code implementations • 15 May 2023 • Zexue He, An Yan, Amilcare Gentili, Julian McAuley, Chun-Nan Hsu
Based on our analysis, we define a disambiguation rewriting task to regenerate an input to be unambiguous while preserving information about the original content.
no code implementations • 28 May 2023 • Zexue He, Marco Tulio Ribeiro, Fereshte Khani
Even when aggregate accuracy is high, state-of-the-art NLP models often fail systematically on specific subgroups of data, resulting in unfair outcomes and eroding user trust.
1 code implementation • ICCV 2023 • An Yan, Yu Wang, Yiwu Zhong, chengyu dong, Zexue He, Yujie Lu, William Wang, Jingbo Shang, Julian McAuley
Recent advances in foundation models present new opportunities for interpretable visual recognition -- one can first query Large Language Models (LLMs) to obtain a set of attributes that describe each class, then apply vision-language models to classify images via these attributes.
no code implementations • 4 Oct 2023 • An Yan, Yu Wang, Yiwu Zhong, Zexue He, Petros Karypis, Zihan Wang, chengyu dong, Amilcare Gentili, Chun-Nan Hsu, Jingbo Shang, Julian McAuley
Medical image classification is a critical problem for healthcare, with the potential to alleviate the workload of doctors and facilitate diagnoses of patients.
no code implementations • 15 Oct 2023 • Noveen Sachdeva, Zexue He, Wang-Cheng Kang, Jianmo Ni, Derek Zhiyuan Cheng, Julian McAuley
We study data distillation for auto-regressive machine learning tasks, where the input and output have a strict left-to-right causal structure.
no code implementations • 21 Oct 2023 • Zexue He, Yu Wang, An Yan, Yao Liu, Eric Y. Chang, Amilcare Gentili, Julian McAuley, Chun-Nan Hsu
Curated datasets for healthcare are often limited due to the need of human annotations from experts.
1 code implementation • 17 Dec 2023 • Yu Wang, Zexue He, Zhankui He, Hao Xu, Julian McAuley
This fine-tuning allows the model to generate explanations that convey the compatibility relationships between items.
no code implementations • 23 Jan 2024 • Jiarui Jin, Zexue He, Mengyue Yang, Weinan Zhang, Yong Yu, Jun Wang, Julian McAuley
Subsequently, we minimize the mutual information between the observation estimation and the relevance estimation conditioned on the input features.
1 code implementation • 19 Feb 2024 • Yu Wang, Zeyuan Zhang, Julian McAuley, Zexue He
To address this issue, we propose Long Video Chat (LVChat), where Frame-Scalable Encoding (FSE) is introduced to dynamically adjust the number of embeddings in alignment with the duration of the video to ensure long videos are not overly compressed into a few embeddings.
no code implementations • 21 Feb 2024 • Zexue He, Leonid Karlinsky, Donghyun Kim, Julian McAuley, Dmitry Krotov, Rogerio Feris
Large Language Models (LLMs) struggle to handle long input sequences due to high memory and runtime costs.