Yinan Zhou

Machine Learning Developer in Speech/Music/Biosignal

About Me

Hi, I’m Yinan. I am interested in how data feminism helps promote a fair and inclusive music industry in terms of gender, race, sexuality, and all the other factors that together influence each person’s experience and opportunities in the world. I am currently seeking job opportunities as a machine learning developer or a data scientist.

Research Interest:

  • Natural Language Processing
  • Data Feminism

Skills:

  • Python (TensorFlow/PyTorch), Django/Flask, AWS, Docker
  • Matlab, C++, C
  • JavaScript/TypeScript, HTML/CSS, React, SQL

Work Experience

DDMAL

Research Assistant

May 2021 - Present

https://ddmal.music.mcgill.ca/
  • Led a team in the development of a music score editor Neon with MEI/MusicXML. Implemented front-end interface with TypeScript, editor toolkit customization to Verovio library in C++, database with IndexedDB, E2E tests with Cypress, resulting in more than 50% reduction in correcting time per folio.
  • Enhanced front-end functionality of an Optical Music Recognition website Rodan with JavaScript and Marionette, and resolved issues in the MEI encoding app using Python.
  • Contributed to building the Virtual Instrument Museum website. Designed and implemented the front-end interface with JavaScript and Bootstrap, back-end with Django, and container with Docker, ensuring a user-friendly and responsive interface.
  • Built an MEI mapping website Cress with TypeScript and deployed the website with webpack and GitHub Pages.

EERS Global Technologies Inc.

Machine Learning Developer

March 2022 - March 2023

https://www.eers.ca/
  • Configured optimal AWS SageMaker settings, and documented infrastructure architecture.
  • Developed an automated pipeline for audio preprocessing, data labelling, and feature extraction in Python to create an unstructured dataset, generated visualizations, and reported insights on spectral characteristics.
  • Built Machine Learning models for breathing rate estimation, speech intelligibility enhancement, and out-of-ear detection, and created pipelines in Python and MATLAB for automated generation of tables and figures, facilitating thorough model comparisons on various aspects.
  • Authored comprehensive deliverables and worked closely with real-time developer to REAPER plugin prototyping and with hardware developer for product prototyping.

Education

McGill University

Master of Arts

2020 - 2022

  • Music Technology, Schulich School of Music
  • GPA: 4.00
  • Supervisor: Prof. Ichiro Fujinaga
  • CIRMMT Student Member
  • DDMAL Research Assistant
  • Thesis: Music Emotion Recognition on Lyrics Using Natural Language Processing

Communication University of China

Bachelor of Engineering

2016 - 2020

  • Network Engineering, Information Engineering School, Faculty of Information Science and Technology
  • GPA: 3.67/4.00, Average Score: 89/100
  • 2016 Excellent Graduate Thesis: Research on Optical Music Notation Recognition Based on Convolutional Neural Network

Projects

Data Scientist Salary Estimation Model

  • Scraped data scientist opening information within Canada from glassdoor.ca using Selenium, and extracted features and built a dataset.
  • Conducted Exploratory Data Analysis (EDA) and created data visualizations with Seaborn to assess market demand for data-related positions, and identify essential skills for professional development strategies.
  • Built machine learning models with custom loss function to estimate data scientist salary, and packaged it into a client facing API using Flask.

Data Analysis: Rum Ratings Website

  • Scraped rumratings.com using BeautifulSoup and saq.com using Selenium to construct a comprehensive rum dataset, and performed data cleaning and preprocessing.
  • Conducted an in-depth analysis of the rum market, identified market gaps, and predicted potential opportunities for specific rum varieties to meet consumer demand.
  • Developed a web-based interactive dashboard using Plotly and Dash to visualize the findings.

AI Song Contest - Poetry-based Chinese Traditional Music Generation Using Transfer Learning

  • Built a new dataset of 6,500 ancient Chinese poetries and corresponding meolodies using OMR techniques, whose lyrics are poetries from more recent times.
  • Finetuned the template-to-melody module in the Telemelody model with the new dataset.
  • Regenerated melodies using the finetuned model to restore the melody of poetry whose melody has been lost.

MUMT 621 Multimodal Music Emotion Recognition Using Convolutional Neural Network

This project explores the performance of multimodal learning on Music Emotion Recognition. Unimodal and bimodal methods on audio and lyrical features were built and compared. Both middle fusion and late fusion were applied for bimodal methods. This result demonstrates that the combination of two feature domains can improve the MER performance.

MUMT 501 Dereverberation on Music Signals

This project implements a block-based dereverberation algorithm. The hyperparameter settings are explored on a singing voice recording. Moreover, the algorithm is also implemented on a short clip of a symphonic music piece by Mozart. It is found that this algorithm has limited effect on signals with high reverberation. The hyperparameter setting for high reverberation extraction is also discussed.

MUMT 618 Realization and Comparison on Ukulele Sound Synthesis Models

This project implements both Karplus-Strong algorithm and digital waveguide model to synthesize ukulele sound. A system is then built to play MIDI files with the resulting sound of two models.

MUMT 605 Cepstral Analysis of Audio Signals

This project introduces the basic concepts of cepstrum and cepstrogram. Their implementations in pitch estimation, spectral envelope estimation, and audio feature extraction are also covered.

Research Experience

College Students Innovation and Entrepreneurship Training Program in Automatic Classification of Music Emotion Based on Machine Learning

Team Leader

April 2018 - April 2019

  • Used web crawler to collect MIDI files on YouTube with music emotion labels.
  • Extracted MFCC, chords, and other features from music fragments.
  • Chose and constructed appropriate neural networks.
  • Trained the model to get the output value; calculated error and updated model parameters to reduce error.

Scientific Research Camp on School Bullying Detection Based on Pattern Recognition at Harbin Institute of Technology

August 2018

  • Preprocessed motion and speech data using techniques including wavelet filtering, pre-emphasis, windowing, and Fourier transform.
  • Performed feature extraction, filtering, and dimensionality reduction on motion and speech data using quartile box charts and PCA algorithms.
  • Classified data using KNN, GMM, and decision trees.

Publications

  • Pang, Long, Ao Li, Yinan Zhou, Chen Yang, Yizhuang Xie, and He Chen. “Word length Optimization Method for Radix-2 k Fixed-Point Pipeline FFT Processors.” In 2019 IEEE International Conference on Signal, Information and Data Processing (ICSIDP), pp. 1-4. IEEE, 2019.
  • Zhao, Wei, Yinan Zhou, Yun Tie, and Yushu Zhao. “Recurrent Neural Network for MIDI Music Emotion Classification.” In 2018 IEEE 3rd Advanced Information Technology, Electronic and Automation Control Conference (IAEAC), pp. 2596-2600. IEEE, 2018.
  • Zhao, Wei, Wang, Lihao, Huang, Jingwen, Zhou, Yinan. “Music Emotion Recognition Based on Feed-Forward Neural Network.” Communication University of China Journal – Natural Science Edition. Vol. 25, No. 4 (2018): 1-5. Grade: 98.88%, issue date: 2021.05

Intern Experience

Heilongjiang Broadcast TV Station

Audio and Video Engineer

July 2019 - August 2019

  • Responsible for the audio and video recording of eight TV programs, including Happy Poker Party and Essence of Longjiang.
  • Responsible for the audio and video during three live broadcasts, including Weekend Tour of Longjiang and Golden Dreams.
  • Participated in the “Heilongjiang Province Celebration of the 70th Anniversary of the Founding of New China” press conference and live broadcast.

Jilin People’s Broadcasting Station

Audio Assistant

August 2018

  • Installed new orchestration system for the recording system.
  • Debugged recording system and renovated the recording studio.
  • Arranged and recorded the original song “Happy Village Convention.”
  • Transcribed the song “The Best Stage.”

Communication University of China Radio Station

Audio Editor

March 2017 - July 2018

  • Recorded, edited, and broadcast radio programs.
  • Synchronized all programs to the online platform.
  • Produced 38 radio programs with an average duration of 20 minutes each.

Robotics Team, Communication University of China

Team Leader

July 2017 - July 2019

  • Apply techniques and concepts from STM 32 microcontrollers and Arduino SCM to build robots based on competition challenges.
  • Participate in robotics competitions in the five provinces of North China.
  • Assign tasks to team members and schedule competition preparation sessions.
  • Participate in school smart tracking car competition.

Coursera Certification

Awards and Scholarships

  • Schulich School of Music
    • Graduate Excellence Fellowship Awards
  • Communication University of China, Beijing, China
    • 2018 First-Class Scholarship, awarded for ranking 2nd in the class
    • 2018 Individual Scholarship for Academic Improvement, awarded by the Information Engineering School, Faculty of Science and Technology
    • 2018 First Prize, 3rd Annual Science and Technology Innovation Competition Robot Match
    • 2017 Third-Class Scholarship, awarded for ranking 6th in the class
    • 2017 Three-Merits Student Award
  • Beijing College Students Sound Trend Meter “College Broadcast Unit” Genre Broadcast, Beijing, China
    • 2018 Excellent Works Award
    • 2017 Second Prize
  • CTV Guangxin Innovation Scholarship, Beijing, China
    • 2018 Excellence in Innovation Award for paper “Recurrent neural network for MIDI music emotion classification”
  • Gaojiao Cup National College Students Mathematical Modeling Contest, Beijing, China
    • 2018 Group A Second Prize
    • 2017 Group A Successful Participation Award
  • Dance Robot Competition and College Student Robot Competition
    • 2017 Group Second Prize, Beijing, China
    • 2017 Group Second Prize, North China

Volunteer Experience

AIESEC in CUC

Manager of Overseas Internship Project Department (Outgoing Global Entrepreneur & Talent)

September 2017 - December 2017

  • Led team members to plan promotional flyers and literature, as well as a new media campaign to promote overseas internships.
  • Assisted applicants in contacting overseas branches and applying for relevant visas.
  • Planned AIESEC in CUC (Communication University of China) gatherings, including poster design, invitation design and mailing, and party favors.

AIESEC in AAST in Cairo

Global Volunteer

August 2017 - September 2017

  • Promoted tourism and economic development in Egypt on social media platforms with the aim of reducing prejudices against Egypt.
  • Joined with 20 volunteers from multiple countries to visit ten cities in Egypt.
  • Researched Egyptian life, culture, religion, and history.

AIESEC in CUC

Manager of Talent Development Department

January 2017 - September 2017

  • Responsible for the new recruits for Spring and Autumn 2017
  • Coordinated social media campaigns, poster design, and marketing copy to promote AIESEC.
  • Performed market research on potential volunteers.
  • Performed candidate consultation and interviewing.
  • Designed performance evaluation criteria for team members.

AIESEC in CUC

Member of International Youth Immigration Exchange Department

June 2016 - December 2016

  • Recruited 7 new members to the AIESEC local volunteer team.
  • Planned the Impression Beijing public welfare program to protect traditional Chinese culture by interviewing artists, making a video, and making traditional artwork with foreign volunteers.