余福浩 Fu-Hao Yu

[email protected], +886 980-152-388

My Master's research is about deep machine learning and natural language processing (NLP) tasks such as automatic speech recognition (ASR), language modeling (LM), text classification, etc. I like learning new technology, such as machine learning and cloud computing applications. I can solve the problem independently and have excellent communication and collaboration skills.

Education

National Taiwan University of Science and Technology

M.S. in Computer Science and Information Engineering

  • GPA: 4.08/4.3
  • Research: Speech Recognition, Deep Learning, Natural Language Processing
  • Master's thesis: Non-autoregressive Transformer-based End-to-end ASR using Pre-trained Language Models
  • Outstanding teaching assistant award: Probability and Statistics

2019 - 2021

National Taiwan University of Science and Technology

B.S. in Computer Science and Information Engineering

  • GPA: 3.93/4.0
  • Academic excellence award (4 times)
  • Honorary member of the Phi Tau Phi Scholastic Honor Society

2015 - 2019

Experience

Research project: The application of business conversation using meta-learning technology

  • Researched code-switching ASR using meta-learning technology.
  • Implemented ASR model based on MAML algorithm using Pytorch.

May - Sep 2020

Developer • Intelligent Ironman Creativity Contest 

  • Developed and maintained the official website and registration system.
  • Introduced Cloudflare DNS service and Google analytics.
  • Created and implemented a creative task using sensors.

Sep 2019 - Aug 2020

Technical Support • Intelligent Ironman Creativity Contest

  • Maintained the web-based contest system.
  • Hosted Arduino tutorial for more than 150 contestants.
  • Solved software and hardware problems for more than 150 contestants.

July-Aug 2019 & July-Aug 2018

Awards

Speech Recognition Classification Competition - Third Prize • AIdea

  • Implemented speech recognition methodology and text classification model using pre-trained language models.

2020

The 44th National Skills Competition - IT Software Solutions for Business - Silver Medal Award • Ministry of Labor

  • Developed and designed Windows forms applications (.NET Framework) using C#.
  • Created relational database using MS SQL.

2014

Projects

Non-autoregressive Transformer-based End-to-end ASR using Pre-trained Language Models

  • Proposed a novel transformer-based non-autoregressive end-to-end ASR model based on BERT.
  • Achieved competitive or superior results on the Chinese AISHELL-1 corpus when compared to well-practiced and/or state-of-the-art ASR models.

Automatic Speech Recognition Models

  • Implemented famous end-to-end ASR models based on RNN or Transformer architecture such as LAS and LASO using Pytorch.

Creativity Tasks of Intelligent Ironman Creativity Contest

  • Created and implemented a creative task using ultrasound sensors, seven-segment display, and keyboard modules.
  • Developed a creative task using C# programming.

Skills

   Python     C#      C      C++      HTML     JavaScript     PHP

  My SQL     MS SQL     Git

Speech Recognition, Deep Learning, Machine Learning, and Natural Language Processing

  • Implemented neural network model using Pytorch framework.
  • Developed E2E ASR model using speech processing toolkit (e.g., Espnet).
  • Fine-tuned pre-trained language models (e.g., BERT) for downstream tasks.


Powered by CakeResumePowered by CakeResume