Wenxiang Jiao (焦文祥)   Senior Researcher NLP Center Tencent AI Lab Shenzhen, China E-mail: wenxiangjiaonju@gmail.com Google Scholar GitHub

🔥 News 🔥

2022-09-24: Won 1st place (team: Borderline) at WMT'2022 Large-Scale Machine Translation Evaluation for African Languages (Constrained Track).

2022-04-26: Self-developed Chinese<=>Japanese translation systems launched on Tencent TranSmart.

2022-02-24: One paper accepted to ACL 2022.

2022-02-01: One paper accepted to IEEE Transactions on Audio, Speech and Language Processing.

2021-11-02: Joined Tencent AI Lab as a Senior Researcher.

2021-09-15: Our project on "Exploiting Large-Scale Data for NMT" won the Technology Innovation Award (top 10%) of 2020 Tencent AI Lab Rhino Bird Project!

2021-08-26: Just passed my Ph.D. oral defense. Many thanks to my supervisors and committees!

Biography

I'm now working at Tencent AI Lab, focusing on researches regarding multilingual NMT and pretraining for NMT. Before that, I just received my Ph.D degree from The Chinese University of Hong Kong, under the supervision by Prof. Irwin King and Prof. Michael R. Lyu. My Bachelor degree and Mphil degree were both achieved at Nanjing University in 2015 and 2017, respectively.
The time line of my research is presented as below:

• 2021-Now: Improving multilingual pretraining models (ACL'22) and multilingual NMT.
• 2019-2021: Exploiting parallel data (EMNLP'20; TASLP'22) and monolingual data (ACL'21) for NMT by data augmentation.
• 2017-2019: Hierachical modeling (NAACL'19; AAAI'20) and pretraining (EMNLP'20 Findings) for emotion recognition in conversations.

Interns

I have benefited a lot from working with these excellent interns.

• Mar. 2022 - Now
  Jen-tse Huang: Chinese University of Hong Kong
  Directions: Robustness in multilingual pretrained models
  
• Aug. 2021 - Now
  Yifan Hou: ETH Zurich
  Directions: Multilingual pretrained LMs, multilingual KG
  
• Aug. 2021 - Now
  Wenxuan Wang: Chinese University of Hong Kong
  Directions: Pretraining for NMT, multilingual NMT
  
• Aug. 2021 - Aug. 2022
  Jiarui Li: Chinese University of Hong Kong (Shenzhen)
  Directions: Adapting multilingual MT and MAE into THUMT
  

Publications

Exploiting Inactive Examples for Natural Language Generation with Data Rejuvenation
Wenxiang Jiao, Xing Wang, Shilin He, Zhaopeng Tu, Irwin King, Michael R. Lyu
IEEE/ACM Transactions on Audio, Speech, and Language Processing. 2022.

Understanding and Improving Sequence-to-Sequence Pretraining for Neural Machine Translation
Wenxuan Wang, Wenxiang Jiao, Yongchang Hao, Xing Wang, Shuming Shi, Zhaopeng Tu, Michael R. Lyu
ACL 2022

Self-training Sampling with Monolingual Data Uncertainty for Neural Machine Translation
Wenxiang Jiao, Xing Wang, Zhaopeng Tu, Shuming Shi, Michael R. Lyu, Irwin King

[code]
ACL 2021

Data Rejuvenation: Exploiting Inactive Training Examples for Neural Machine Translation
Wenxiang Jiao, Xing Wang, Shilin He, Irwin King, Michael R. Lyu, Zhaopeng Tu

[code]
EMNLP 2020

Exploiting Unsupervised Data for Emotion Recognition in Conversations
Wenxiang Jiao, Michael R. Lyu, Irwin King

[code]
EMNLP 2020 Findings (short)

Real-Time Emotion Recognition via Attention Gated Hierarchical Memory Network
Wenxiang Jiao, Michael R. Lyu, Irwin King

[code]
AAAI 2020

HiGRU: Hierarchical Gated Recurrent Units for Utterance-Level Emotion Recognition
Wenxiang Jiao, Haiqin Yang, Irwin King, Michael R. Lyu

[code]
NAACL-HLT 2019

Multi-Task Learning with Shared Encoder for Non-Autoregressive Machine Translation
Yongchang Hao, Shilin He, Wenxiang Jiao, Zhaopeng Tu, Michael Lyu, Xing Wang
NAACL-HLT 2021

Improving Word Representations: A Sub-sampled Unigram Distribution for Negative Sampling
Wenxiang Jiao, Irwin King, Michael R. Lyu
arXiv 2019

Academic Activity

2022
Workshop speech at NLPCC 2022
Reviewer of ACL/ARR 2022

2021
Reviewer of the Neurocomputing Journal
Reviewer of ACL 2021

2020
Research talk at AI TIME on Data Rejuvenation
Reviewer of ACL 2020, AAAI 2021
Secondary reviewer of EMNLP 2020, COLING 2020, EACL 2020

Experiences

• Oct. 2019 - Dec. 2019
  Research Intern: Tencent AI Lab
  Mentor: Xing Wang, Zhaopeng Tu
  Worked on data analysis and exploitation in neural machine translation.
  Published one paper to EMNLP 2020, ACL 2021 and TASLP 2022, respectively.

Paper Reading

Jun. 2021
Multi-Task Learning for Multilingual Neural Machine Translation (ACL 2020-2021) [slides]

Nov. 2020
Exploiting Monolingual Data at Scale for Neural Machine Translation (EMNLP 2019) [slides]

Aug. 2020
A Closer Look at Memorization in Deep Networks (ICML 2017) [slides]

Jul. 2020
Towards Explaining the Regularization Effect of Initial Large Learning Rate in Training Neural Networks (NeurIPS 2019) [slides]

Education

Shatin, NT, Hong Kong
The Chinese University of Hong Kong (CUHK)

Aug. 2017 - Aug. 2021
Doctor of Philosophy Student, Computer Science & Engineering
Advisor: Irwin King (IEEE Fellow), and Michael R. Lyu (IEEE Fellow, ACM Fellow)

Nanjing, China
Nanjing University (NJU)

Sep. 2015 - Jun. 2017
Mphil of Engineering, Optical Engineering
Advisor: Guanghui Wang

Nanjing, China
Nanjing University (NJU)

Sep. 2011 - Jun. 2015
Bachelor of Engineering, Information Engineering

Awards

Tencent, 2022
1st place at WMT'2022 Large-Scale Machine Translation Evaluation for African Languages (Constrained Track)

CUHK, 2021
Technology Innovation Award (top 10%) of 2020 Tencent AI Lab Rhino Bird Project

CUHK, 2017 - 2021
Full Postgraduate Studentship

NJU, 2016
Huawei Scholarship

NJU, 2015
Outstanding Graduates Award

NJU, 2014
Bejing Bank Outstanding Scholarship

Ministry of Education of China, 2013
China National Scholarship

COMAP, 2014
Honorable Mentions in American Mathematical Contest in Modeling (MCM)

CSIAM, 2014
Honorable Mentions in National College Mathematical Contest in Modeling

COS, 2014
Honorable Mentions in National College Optoelectronic Design Competition

Skills

Programming Languages: Python, C/C++, Matlab

Deep Learning Frameworks: Pytorch, Tensorflow

Language

TOEFL: 100

GRE: V154 + Q170 + AW3.5