User profiles for Kai-Wei Chang

Kai-Wei Chang

Associate Professor, UCLA
Verified email at kwchang.net
Cited by 32656

Man is to computer programmer as woman is to homemaker? debiasing word embeddings

T Bolukbasi, KW Chang, JY Zou… - Advances in neural …, 2016 - proceedings.neurips.cc
The blind application of machine learning runs the risk of amplifying biases present in data.
Such a danger is facing us with word embedding, a popular framework to represent text data …

[PDF][PDF] LIBLINEAR: A library for large linear classification

RE Fan, KW Chang, CJ Hsieh, XR Wang… - the Journal of machine …, 2008 - jmlr.org
LIBLINEAR is an open source library for large-scale linear classification. It supports logistic
regression and linear support vector machines. We provide easy-to-use command-line tools …

Generating natural language adversarial examples

…, A Elgohary, BJ Ho, M Srivastava, KW Chang - arXiv preprint arXiv …, 2018 - arxiv.org
Deep neural networks (DNNs) are vulnerable to adversarial examples, perturbations to
correctly classified examples which can cause the model to misclassify. In the image domain, …

Visualbert: A simple and performant baseline for vision and language

…, M Yatskar, D Yin, CJ Hsieh, KW Chang - arXiv preprint arXiv …, 2019 - arxiv.org
We propose VisualBERT, a simple and flexible framework for modeling a broad range of
vision-and-language tasks. VisualBERT consists of a stack of Transformer layers that implicitly …

Grounded language-image pre-training

…, L Zhang, JN Hwang, KW Chang… - Proceedings of the …, 2022 - openaccess.thecvf.com
This paper presents a grounded language-image pre-training (GLIP) model for learning
object-level, language-aware, and semantic-rich visual representations. GLIP unifies object …

Learn to explain: Multimodal reasoning via thought chains for science question answering

…, S Mishra, T Xia, L Qiu, KW Chang… - Advances in …, 2022 - proceedings.neurips.cc
When answering a question, humans utilize the information available across different
modalities to synthesize a consistent and complete chain of thought (CoT). This process is …

Chameleon: Plug-and-play compositional reasoning with large language models

…, H Cheng, M Galley, KW Chang… - Advances in …, 2024 - proceedings.neurips.cc
Large language models (LLMs) have achieved remarkable progress in solving various
natural language processing tasks due to emergent reasoning abilities. However, LLMs have …

A dual coordinate descent method for large-scale linear SVM

CJ Hsieh, KW Chang, CJ Lin, SS Keerthi… - Proceedings of the 25th …, 2008 - dl.acm.org
In many applications, data appear with a huge number of instances as well as features. Linear
Support Vector Machines (SVM) is one of the most popular tools to deal with such large-…

Unified pre-training for program understanding and generation

…, S Chakraborty, B Ray, KW Chang - arXiv preprint arXiv …, 2021 - arxiv.org
Code summarization and generation empower conversion between programming language
(PL) and natural language (NL), while code translation avails the migration of legacy code …

Men also like shopping: Reducing gender bias amplification using corpus-level constraints

…, T Wang, M Yatskar, V Ordonez, KW Chang - arXiv preprint arXiv …, 2017 - arxiv.org
Language is increasingly being used to define rich visual recognition problems with supporting
image collections sourced from the web. Structured prediction models are used in these …