Haoran (Felix) Xu

Haoran (Felix) Xu

Senior Researcher

Microsoft GenAI

About Me (徐浩然)

I am a Senior Researcher at Microsoft GenAI. I earned my Ph.D. in Computer Science from Johns Hopkins University in 2024, co-advised by Philipp Koehn and Kenton Murray. My most recent research focuses on foundational training of large language models, machine translation, and multilinguality. Please find up-to-date list of all my publications on my Google Scholar profile.

I also had good fortune to intern at Microsoft Research, Meta (Facebook) AI Research and Amazon Alexa AI.

Interests

  • Large Language Model
  • Machine Translation
  • Multilinguality

Education

  • Ph.D. in Computer Science, 2024

    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

 
 
 
 
 

Senior Researcher

Microsoft GenAI

Oct 2024 – Present Redmond
Phi pre-training and post-training
 
 
 
 
 

Research Scientist Intern

Microsoft Research

May 2023 – Aug 2024 Redmond
Investigated multilingual machine translation for Large Language Models: ALMA Models!
 
 
 
 
 

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

Preprints

Last updated on October 2024

Streaming Sequence Transduction through Dynamic Compression

Last updated on October 2023

Efficiently Harnessing Parameter Importance for Better Training

Publications

October 2024 ICLR 2025 (Spotlight)

X-ALMA: Plug & Play Modules and Adaptive Rejection for Quality Translation at Scale

PDF Code Dataset
October 2024 NAACL 2024

Upsample or Upweight? Balanced Training on Heavily Imbalanced Datasets

PDF
May 2024 ICML 2024

Contrastive Preference Optimization: Pushing the Boundaries of LLM Performance in Machine Translation

PDF Code Dataset Slides
September 2023 ICLR 2024

A Paradigm Shift in Machine Translation: Boosting Translation Performance of Large Language Models

PDF Code Dataset Slides Video
September 2023 ICLR 2024

Error Norm Truncation: Robust Training in the Presence of Data Noise for Text Generation Models

PDF Code
May 2023 EMNLP 2023

Condensing Multilingual Knowledge with Lightweight Language-Specific Modules

PDF Code
May 2023 Findings of EMNLP 2023

Towards Being Parameter-Efficient: A Stratified Sparsely Activated Transformer with Dynamic Capacity

PDF Code
February 2023 Findings of EACL 2023

Language-Aware Multilingual Machine Translation with Self-Supervised Learning

PDF Code
October 2022 EMNLP 2022

The Importance of Being Parameters: An Intra-Distillation Method for Serious Gains

PDF Code Slides
April 2022 Findings of NAACL 2022

Por Qué Não Utiliser Alla Språk? Mixed Training with Gradient Optimization in Few-Shot Cross-Lingual Transfer

PDF Code Poster
See all publications

Contact