Kai Zhao

kzhao.hf [at] gmail.com

Biography

I currently work at Google DeepMind as a research scientist, focusing on Gemini Deep Research.

Previously, I worked on Gemini tool use in the Gemini Extensions team, and neural semantic parsing in the Google Assistant team.

I did my Ph.D. at School of Electrical Engineering and Computer Science, Oregon State University, where my research focused on algorithms and theory in Natural Language Processing (NLP), especially in structured prediction problems including syntactic/semantic parsing, machine translation, and textual entailment. My advisor is Prof. Liang Huang. Before that, I received my Bachelor of Engineering (B.Eng.) degree from Computer Science Department, University of Science and Technology of China (USTC).

Publications

My publication list is also available at Google Scholar.

Liang Huang, He Zhang, Dezhong Deng, Kai Zhao, Kaibo Liu, David Hendrix, and David Mathews. LinearFold: linear-time approximate RNA folding by 5’-to-3’ dynamic programming and beam search. In Proceedings of ISMB 2019.
Kai Zhao and Liang Huang. Joint Syntacto-Discourse Parsing and the Syntacto-Discourse Treebank. In Proceedings of EMNLP 2017 (arXiv, code).
Liang Huang, Kai Zhao, and Mingbo Ma. When to Finish? Optimal Beam Search for Neural Text Generation (modulo beam size). In Proceedings of EMNLP 2017.
Mingbo Ma, Kai Zhao, Liang Huang, Bing Xiang, and Bowen Zhou, Jointly Trained Sequential Labeling and Classification by Sparse Attention Neural Networks. In Proceedings of Interspeech 2017 (arXiv).
Kai Zhao, Liang Huang, and Mingbo Ma. Textual Entailment with Structured Attentions and Composition. In Proceedings of COLING 2016 (arXiv, code, slides).
Feifei Zhai, Liang Huang, and Kai Zhao. Search-aware Tuning for Hierarchical Phrase-based Decoding. In Proceedings of EMNLP 2015.
Kai Zhao, Hany Hassan, and Michael Auli. Learning Translation Models from Monolingual Continuous Representations. In Proceedings of NAACL 2015 (slides).
Kai Zhao and Liang Huang. Type-driven Incremental Semantic Parsing with Polymorphism. In Proceedings of NAACL 2015 (code, slides).
Kai Zhao, Liang Huang, Haitao Mi, and Abe Ittycheriah. Hierarchical MT Training using Max-Violation Perceptron. In Proceedings of ACL 2014 (slides).
Kai Zhao, James Cross, and Liang Huang. Optimal Incremental Parsing via Best-First Dynamic Programming. In Proceedings of EMNLP 2013.
Heng Yu, Liang Huang, Haitao Mi, and Kai Zhao. Max-Violation Perceptron and Forced Decoding for Scalable MT Training. In Proceedings of EMNLP 2013.
Hao Zhang, Liang Huang, Kai Zhao, and Ryan McDonald. Online Learning with Inexact Hypergraph Search. In Proceedings of EMNLP 2013.
Yoav Goldberg, Kai Zhao, and Liang Huang. Efficient Implementation for Beam Search Incremental Parsers. In Proceedings of ACL 2013.
Kai Zhao and Liang Huang. Minibatch and Parallelization for Online Large Margin Structured Learning. In Proceedings of NAACL 2013.

We have a tutorial on Structured Prediction with Perceptron presented in ACL 2014 (slides), and MASC-SLL 2015 (revised slides).

Softwares

Code for some of the papers listed above:

Codes for Fun

Dependency Tree Viewer: a tool to visualize (unlabeled) dependency trees with JavaScript. It supports nested brackets format, CONLL-X format, and Stanford format. (For syntax tree there is a great viewer from Miles Shang.)