Haoran (Felix) Xu

Haoran (Felix) Xu

NLPer

Johns Hopkins University

About Me

I am a Ph.D. student in Computer Science at Johns Hopkins University, affiliated with the Center for Language and Speech Processing (CLSP), co-advised by Philipp Koehn and Kenton Murray. My most recent research focuses on Cross-lingual Learning and Machine Translation.

I also had good fortune to intern at Meta (Facebook) AI Research and Amazon Alexa AI. I will be joining Microsoft Research in the upcoming summer!

Interests

  • Machine Translation
  • Cross-Lingual Learning
  • Multilinguality
  • Pre-Training

Education

  • Ph.D. in Computer Science, Present

    Johns Hopkins University

  • M.S. in Computer Science, 2020

    Johns Hopkins University

  • B.E. in Information Engineering, 2018

    East China University of Science and Technology

Experience

 
 
 
 
 

Research Scientist Intern

Microsoft Research

May 2023 – Aug 2023 Bellevue
Multilingual Machine Translation
 
 
 
 
 

Research Scientist Intern

Meta (Facebook) AI Research

May 2022 – Dec 2022 Menlo Park
Studied self-supervised learning method and mixture of experts (MoE) for multilingual machine translation under the No Language Left Behind (NLLB) group
 
 
 
 
 

Applied Scientist Intern

Amazon Alexa AI

May 2021 – Aug 2021 Seattle
Worked on novel style transfer algorithms that transfer the text style while keeps the main semantics
 
 
 
 
 

Research Intern

Center for Language and Speech Processing

May 2020 – Dec 2020 Baltimore
Investigated cross-lingual transfer learning and machine translation

Publications

VAE based Text Style Transfer with Pivot Words Enhancement Learning

Projects

*

Cross-Lingual Contextual Embedding Space Mapping

Proposed a novel method of contextual embedding mapping and revealed the tight relationship of isotropy, isometry and isomorphsim in contextual embedding spaces.

Efficient Quadratic Programming in Wireless Communication

Developed a new algorithm to reduce the peak energy of data transmission based on Unimodular Quadratic Programming.

Gradual Fine-Tuning for Low-Resource Domain Adaptation

Proposed novel data augmentation approach for domain adaptation, which surpasses state-of-the-art performance in Dialogue State Tracking and Event Extraction tasks.

Image expansion with GANs

Built a deep learning method based on GANs to naturally predict and expand the boundaries of incomplete images

Invisible Signature Security System

Designed a human-computer interaction GUI system to recognize signatures of users with high precision, where users sign their names in the air (invisibly).

Zero-Shot Cross-Lingual Dependency Parsing

Investigated a zero-shot approach for dependency parsing by building a multilingual concept-shared semantic space, which achieves state-of-the-art performance.