Paper Title Number 4
Your Name, You. (2024). "Paper Title Number 3." GitHub Journal of Bugs. 1(3).
CV
Education
- New York University
Master of Applied Mathematics, Sep 2023 - Jun 2025
GPA: 3.61/4.00- Joined the NYU VideoLab from Electrical Engineering in Sep 2024 under the advisory of Professor Yao Wang and postdoctoral researcher Nikola Janjušević.
- Researching the construction of interpretable deep-learning models for solving noisy 3D MRI reconstruction problems, forming the basis of my master’s thesis.
- University of Washington
Bachelor of Physics, Sep 2019 - Jun 2023
GPA: 3.73/4.00
Work & Project Experience
- Optimization of Unambiguous Recognition
May 2024 - Present- New York University
- Duties include:
- Optimized scheduling between disk, memory, and GPU by using asynchronous data processing and mounting, increasing GPU utilization from 60% to 95%.
- Modified SwinGPT2 encoder structure to improve encoder inference efficiency by approximately 4 times without significant loss of accuracy.
- Implemented sliding encoder-decoder streaming input and output, making input size of images or audio and output text unrestricted.
- Unambiguous Recognition
Apr 2024 - May 2024- New York University
- Duties include:
- Proposed that unambiguous recognition should not solely use natural language corpus as the training set, and developed a multimodal model using SwinGPT2 structure for end-to-end visual encoding and text decoding.
- Paper https://arxiv.org/abs/2406.17148 and Code https://github.com/RQLuo/MixTeX
- Mined and processed a dataset of 120M tokens of \LaTeX{} text labels and image inputs.
- Trained MixTex for recognizing multilingual text mixed with formulas and tables.
- Document Dewarping Algorithm
Apr 2022 - Jun 2022- University of Washington
- Duties include:
- Proposed a contour-based mesh subdivision surface flattening method and applied it to scan curved surface pages. A non-neural network approach using traditional graphics methods, suitable for low-resource high-speed processing.
- Six months after open-sourcing, CamScanner launched a similar feature.
- Paper https://arxiv.org/abs/2212.13489 and Code https://github.com/RQ-Luo/contour-dewarp
- Multi-Speaker Voice Synthesis
June 2023 - August 2023- University of Washington
- Duties include:
- Developed a neural network-based multi-speaker text-to-speech system for generating audiobooks.
- Achieved 91% accuracy in identifying the gender, age, and emotion of novel dialogue characters using BERT.
- Mined a 10GB speech dataset, assisted by Whisper for labeling.
- Set up multi-GPU data parallel training on rented servers and completed model training within a week.
- AI Novel Animation Demonstration Video
Publications
Paper Title Number 3
Your Name, You. (2015). "Paper Title Number 3." Journal 1. 1(3).
Paper Title Number 2
Your Name, You. (2010). "Paper Title Number 2." Journal 1. 1(2).
Paper Title Number 1
Your Name, You. (2009). "Paper Title Number 1." Journal 1. 1(1).
Talks
Conference Proceeding talk 3 on Relevant Topic in Your Field
Conference proceedings talk at Testing Institute of America 2014 Annual Conference, Los Angeles, CA
Talk 2 on Relevant Topic in Your Field
Talk at London School of Testing, London, UK
Tutorial 1 on Relevant Topic in Your Field
Tutorial at UC-Berkeley Institute for Testing Science, Berkeley CA, USA
Talk 1 on Relevant Topic in Your Field
Talk at UC San Francisco, Department of Testing, San Francisco, California
Teaching
Service and leadership
- Currently signed in to 43 different slack teams