CV
Education
- M.S. in System Engineering, Beihang University Aug 2020 - July 2022
- Generalist Engineer (= Master of Engineering), Centrale Lille Aug 2018 - July 2020
- B.S. in Mathematics and Applied Mathematics, Beihang University Aug 2015 - July 2018
Research Experiences
Research in natural language generation and pre-trained language models Nov 2020 - Present
Advisor: Wenge RONG, Professor, School of Computer Science and Engineering, Beihang University
- Researched and applied Knowledge graph (KG) embedding methods to generate contextualized natural languages with more details.
- Proposed an attention-based knowledge filter mechanism to filter out irrelevant knowledge.
- Proposed a hierarchical lexicon embedding structure to solve Chinese named entity recognition task and achieved SOTA in 3 commonly used Chinese dataset.
Integrated Large-Scale Model Training System with Efficient Parallelization Techniques Aug 2021 - Present
Advisor: Yang YOU, Professor, School of Computer Science and Engineering, National University of Singapore
- Models and datasets are getting significantly larger, making a single GPU not enough to accommodate the entire model.
- Developed a distributed machine learning framework that combines hybrid parallel training methods, aiming to support the AI community to write distributed models in the same way as how they write models normally.
- Conducted experiments on the developed system, e.g. scaling test/ performance alignment test on ViT-ImageNet.
Straightforward Neural Network Method for Mesh Smoothing Sep 2020 - Nov 2020
Advisor: Yufei GUO, PhD Researcher, X lab
- Proposed a the first neural network approach to solve mesh smoothing problem.
- Achieved efficiency 4 times faster than traditional optimization-based methods while being comparable in terms of mesh quality.
Internship Experiences and Projects
Ainnovation, Department of Industrialized Computer Vision | R&D Intern | Beijing Jul 2020 - Sep 2020
- Built an Intelligent Edge Video Platform providing real-time functional interfaces that are adaptable to various commercial applications in terms of model development and hardware compatibility. (https://www.ainnovation.com/mmo/matrixvision)
- Developed and encapsulated core functions of the platform including video/photo encoding, decoding, cropping, augmenting, etc. on Huawei HiSilicon chip (C++).
- Trained and deployed 2 sub-applications on the platform: online face recognition system and kitchen monitoring system.
Online Chinese Chess Game | WEB course project | Lille June 2020 - June 2020
- Created a game application on web page which permits two players on different web browsers to participate in the same match on live (Documentation).
- Defined the specifications of gameplay, functions and main user interface.
- Developed both front end and back end. Realized the database, registration and sign-in, web page display, as well as game realization.
Drone Project, BL evolution (https://www.bl-evolution.com) | Co-Leader | Lille Dec 2018 - Feb 2020
- Designed and assembled devices (drones) to measure various environmental data from the air.
- Participated in project establishment, budget estimation, schedule arrangement, work allocation, and final product delivery.
- Assembled sensors on Raspberry Pi to monitor/record/store data.
- Processed the observed data, represented the analytical results on a map database, and applied target detection to the images to identify thermal energy, roof ridges, or green coverage.
Using wavelet transform and gradient descent to trait image and audio signals | Lille Feb 2020 - Feb 2020
- Applied Haar wavelet transform and discrete wavelet transform in image processing projects to make images transformed with multi-resolution and multi-scale capability
- Rendered very large gradient values on image edges for object edge detection without training.
- Performed noise removal and image restoration on unclear image sets.
- Applied Haar wavelet transform for noise reduction and audio segmentation