K
Kai-Wei Chang
Researcher at University of California, Los Angeles
Publications - 262
Citations - 23031
Kai-Wei Chang is an academic researcher from University of California, Los Angeles. The author has contributed to research in topics: Computer science & Word embedding. The author has an hindex of 42, co-authored 183 publications receiving 17271 citations. Previous affiliations of Kai-Wei Chang include Boston University & Amazon.com.
Papers
More filters
Proceedings Article
Weight Perturbation as Defense against Adversarial Word Substitutions
Jianhan Xu,Linyang Li,Jiping Zhang,Xiaoqing Zheng,Kai-Wei Chang,Cho-Jui Hsieh,Xuanjing Huang +6 more
Posted Content
Disentangling Semantics and Syntax in Sentence Embeddings with Pre-trained Language Models
TL;DR: The authors disentangle semantics and syntax in sentence embeddings obtained by pre-trained language models, leading to better robustness against syntactic variation on downstream semantic tasks, and outperforms state-of-the-art sentence embedding models on unsupervised semantic similarity tasks.
Journal ArticleDOI
GIVL: Improving Geographical Inclusivity of Vision-Language Models with Pre-Training Methods
TL;DR: GIVL as discussed by the authors is a pre-trained model that learns geo-diverse visual concepts by pre-training Image Knowledge Matching (IKM) and Image Edit Checking (IEC).
Proceedings Article
TAGPRIME: A Unified Framework for Relational Structure Extraction
I-Hung Hsu,Kuan-Hao Huang,Shuning Zhang,Wen-Huang Cheng,Prem Natarajan,Kai-Wei Chang,Nanyun Peng +6 more
TL;DR: This work proposes to take a unified view of all these tasks and introduce TAGPRIME to ad011 dress relational structure extraction problems, and finds that the self-attention words in pre-trained language models contain more information about the given condition, and hence become more suit020 able for extracting specific relationships for the 021 condition.
Posted Content
Learning to Represent Bilingual Dictionaries
TL;DR: The authors propose a neural embedding model that leverages bilingual dictionaries to map the literal word definitions to the cross-lingual target words, for which they explore with different sentence encoding techniques.