Grants
- Primary Investigator, “Identification of growth inhibitors in the algal reused media and its removal”, 2016/01-2017/09, Chinese postdoctoral science funds for the outstanding Post- doctors, $8 000.00.
- Major Investigator, “Membrane fouling mechanisms caused by Algogenic Organic Matter (AOM) and its mitigation during microalgae harvesting process”, Natural Science Foundation of China, 2017/01-2020/12, $ 90, 000.00.
- Major Investigator, “Combining biofuel and high-value bioproducts to meet the renewable fuel standard”, US Department of Energy, 2016/01-2019/12, $ 5, 200, 000.00.
- Major Investigator, “The key technologies in the algal harvesting and downstream process”, State Investigation and Development Corporation, 2015/02-2017/04, $ 400, 000.00.
- Major Investigator, “Study on Cell Aggregation Mechanism and Dynamics in Microalgae Biomass Production”, Natural Science Foundation of China, 2017/01-2019/12,$ 40, 000.00.
- Major Investigator, “Joint Effects of Four Alkaloids on Two Algae and Their Ecological Safety Assessment”, Natural Science Foundation of China, 2012/01-2015/12, $ 40, 000.00.
Projects
- AI against Cancer, Data science Hackathon, 2021-Natural Language Processing System for Retrieving Information about Cancer Phenotypes from Published Articles (AI and Cancer Hackathon).
- Develop a pipeline to identify cancer-related genes and Human Phenotype ontologies (HPO) in given cancer research article by searching similar HPO definition for each relevant sentences of a document.
- Using the a pre-trained deep learning model BioBERT, 1183 papers were analyzed in PubMed, and identified 2 times more genes than that HPO represents.
- Significantly improved the accuracy in screening the relevant genes.
- Object Detection with Mask R-CNN: Drone Images of Transmission Tower Components
- Annotated images using VGG Image Annotator, and parsed the annotation JSON files into a coco-like format for model training, validation and testing.
- Model training: Matterport’s implementation of Mask R-CNN, running on a TensorFlow and Keras backend on Google Colab. mAP score was calculated to evaluate the model performance, and our model showed a mAP score of 0.972.
- Music Genre Recognition with VGG16.
- Established a method to classify music genres using audio spectrograms created from original soundtracks, and convert an voice recognition to well-studied image recognition problem.
- Spectrograms were used as visual representation for feature extraction using by a deep learning model VGG-16 with the Adam optimizer.
- Model performance: an accuracy of > 80% for genre classification in both training and testing.
- Build a Bookstore management system by Java GUI (Source: CS205 OOP)
- Book information was read into ArrayList objects, and visualized by Java swing JTable and JList
- The class to create Jtable was inherited from AbstractTableModel
- Gui inherented JFrame and implements ListSelectionListener, establishing JList, JButton, JTextField and JLabels.
- Book information was read into ArrayList objects, and visualized by Java swing JTable and JList
- Design an online Amazon-like bookstore store website
- Back end: Python-based framework flask was used for main structure. Sqlite and SQLAlchemy were used to build the database. Jinja script was used for communication between backend and frontend.
- Front end: HTML, CSS, Bootstrap, and JavaScript was used to create static website. Inheritance is used simply the code. Functions including user login, sign up and interacting with database were also included.
- Design an Amazeriffic website
- HTML, CSS Bootstrap, and JavaScript was used to create static website.
- JQuery, Node.js, NoSQL and MongoDB were implemented to make the web server hosted in a cloud platform by docker and Kubernetes.
- Implement the PageRank algorithm in Racket
- Wrote a graph algorithm to insert a graph (represented as a list of links), calculate the PageRank score for a given graph, save the score in a hash set, sort the hash set based on its hash value and give the order of nodes.
- Build a usable VPN through Amazon AWS cloud platform and Pritunl client.
- Started a EC2 Oracle Linux virtual machine on AWS, and install pritunl and MangoDB on it.
- Routing on AWS, setup Mango Database, configure server and update security group.
- MAGIC (Marine AlGae Industrialization Consortium) – Combining biofuel and high-value bioproducts to meet the renewable fuel standard (DOE Funded Project #DE-EE0007091, USA). My roles in this project are (1) to harvest 8 strains of 40 kg Ash free dry weight algal biomass for biofuel and feed production; (2) Extract algal oil using wet biomass (2017- 2019).
- Study the key technologies in the downstream processing of microalgae (State Development & Investment Company-funded project, China). My role in this project is to characterize the growth inhibitors in the recycled algal culture media and develop effective technologies for media recycling (2015-2017).
- Investigation on the microbial community structure of constructed wetland in Beijing Olympic Forest Park (Beijing, China). My role in this project was to sample field soil sample from wetland system in summer and winter (2011-2012).