Zhenyun Deng
I finished my PhD thesis from School of Computer Science, University of Auckland in September 2022. I have been fortunate to be advised by Prof. Michael Witbrock and Dr. Patricia Riddle. Before that, I received M.S. in Computer Science from Guangxi Normal University, advised by Prof. Shichao Zhang.
I was a reasearch assistant at Strong AI Lab (SAIL) at UoA. Prior to UoA, I previously worked as an NLP engineer in CNKI, Beijing. See CV for more information.
Email  / 
Google Scholar  / 
Github  / 
|
|
Research Interests
- Interpretable AI, Question Answering, Fact Checking
- Natural Language Understanding/Reasoning, Graph Neural Networks
- Deep Learning Applications in NLP, Continual/Lifelong Learning, Machine Learning, Causal Inference, etc.
|
Experiences
-
Aug 2022 - Feb 2023, University of Auckland, Auckland, New Zealand. Research assistant at
Strong AI Lab
-
Spet 2016 - Sept 2018, China National Knowledge Infrastructure (CNKI, 中国知网), Beijing, China. NLP Engineer at
Big Data Analysis Group
|
Selected Publications
Google Scholar
for all publications. (*: equal contribution)
|
|
Prompt-based Conservation Learning for Multi-hop Question Answering
Zhenyun Deng, Yonghua Zhu, Yang Chen, Qianqian Qi, Michael Witbrock, Patricia Riddle
International Conference on Computational Linguistics (COLING), 2022.
[paper]
We propose the Prompt-based Conservation Learning (PCL) framework for multi-hop QA, which acquires new knowledge from multi-hop QA tasks while conserving old knowledge learned on single-hop QA tasks, mitigating forgetting.
|
|
Interpretable AMR-Based Question Decomposition for Multi-hop Question Answering
Zhenyun Deng, Yonghua Zhu, Yang Chen, Michael Witbrock, Patricia Riddle
International Joint Conference on Artificial Intelligence (IJCAI), 2022.
[paper]
Oral presentation (4%)
We propose a Question Decomposition method based on Abstract Meaning Representation (QDAMR) for multi-hop QA, which achieves interpretable reasoning by decomposing a multi-hop question into simpler subquestions and answering them in order.
|
|
Explicit Graph Reasoning Fusing Knowledge and Contextual Information for Multi-hop Question Answering
Zhenyun Deng, Yonghua Zhu, Qianqian Qi, Michael Witbrock, Patricia Riddle
Deep Learning on Graphs for Natural Language Processing (DLG4NLP), 2022.
[paper]
We describe a structured Knowledge and contextual Information Fusion GNN (KIFGraph) whose explicit multi-hop graph reasoning mimics human step by step reasoning.
|
|
Multi-Step Deductive Reasoning Over Natural Language: An Empirical Study on Out-of-Distribution Generalisation
Qiming Bao, Alex Yuxuan Peng, Tim Hartill, Neset Tan, Zhenyun Deng, Michael Witbrock, Jiamou Liu
Neural-Symbolic Learning and Reasoning (NeSy), 2022.
[paper]
We introduce IMA-GloVe-GA, an iterative neural inference network for multi-step reasoning expressed in natural language.
|
|
Multi-scale Graph Fusion for Co-saliency Detection
Rongyao Hu*, Zhenyun Deng*, Xiaofeng Zhu (co-first author)
AAAI Conference on Artificial Intelligence, 2021.
[paper]
We propose a new co-saliency detection framework which includes two strategies to improve the discriminative ability of the features.
|
Teaching
- Graduate Teaching Assistant, University of Auckland, 2022
   Fundamentals of Database Systems (CS351/SE351/CS751).
- Graduate Teaching Assistant, University of Auckland, 2020
   Fundamentals of Database Systems (CS351/SE351/CS751).
|
Awards
- PhD Thesis Writing Award, 2022
- Computer Science Graduate Student Travel Fund, 2022
- New Zealand Tertiary Education Commission Fund, 2022
- PhD Research Project Scholarship, 2018
- China National Scholarship, 2015
- Graduate Student Full Scholarship, 2013
|
Academic Services
Conference Reviewer: ACL/EMNLP/NAACL Area Chair, IJCAI, AAAI, ACMMM, CIKM 2022, ACL Rolling Review (ARR)
Journal Reviewer: Neurocomputing
|
Template Credit: Jon Barron
|