Research Scientist

• Jan 2023 - Present

Research Assistant

• Aug 2019 - Present

Object Detection and Pedestrian Navigation (Python/Java, PyTorch): Developed an augmented reality Android prototype that uses landmarks as beacons to navigate users in urban areas. Landmarks are detected using deep-learning-based models implemented in PyTorch. Synthetic Data Generation (Python, Tensorflow): Proposed a deep learning model that applies Long Short-Term Memory networks (LSTM) and Generative Adversarial Networks (GANs) to generate privacy-preserving synthetic data. Recommendation via Federated Learning (Python, PySyft): Used Federated Learning and Additive Secret Sharing to build a privacy-preserving recommendation system for Point-of-Interest (POI) recommendation. Data Privacy Techniques (Python/R): Worked with data privacy methods such as de-identification, K-anonymity, differential privacy; explored the effectiveness of location privacy techniques (e.g., aggregation, geomasking) on massive geotagged social media data. Show less

Co-Founder

• Sep 2019 - Present

GISphere is a voluntary project started by a group of Chinese young scholars who study GIScience worldwide. GISphere is devoted to bridging the digital and information gap for Chinese undergraduates who major in geography-related majors and help them to access the most up-to-date information on master/Ph.D. opportunities, funding availabilities, webinars, and competitions globally. So far (Dec 2020) more than 700 students have been benefited from the program, free of charge. GISphere is a voluntary project started by a group of Chinese young scholars who study GIScience worldwide. GISphere is devoted to bridging the digital and information gap for Chinese undergraduates who major in geography-related majors and help them to access the most up-to-date information on master/Ph.D. opportunities, funding availabilities, webinars, and competitions globally. So far (Dec 2020) more than 700 students have been benefited from the program, free of charge.

Research Scientist

• Sep 2022 - Jan 2023

Research Scientist

• Sep 2022 - Jan 2023

AI Resident

• May 2021 - May 2022

AI Resident

• May 2021 - May 2022

Geospatial Vision Intern

• Jun 2020 - Aug 2020

Geospatial Vision Intern at Sturfee Inc. Designed a computer vision algorithm for projecting image from ground view to satellite view (Python, Numpy, OpenCV, OpenGL); Designed an algorithm for automatic feature matching between ground view image and satellite image (Python, Numpy, OpenCV, OpenGL); Ground feature filter using semantic segmentation on ground view images (Python, Tensorflow). Geospatial Vision Intern at Sturfee Inc. Designed a computer vision algorithm for projecting image from ground view to satellite view (Python, Numpy, OpenCV, OpenGL); Designed an algorithm for automatic feature matching between ground view image and satellite image (Python, Numpy, OpenCV, OpenGL); Ground feature filter using semantic segmentation on ground view images (Python, Tensorflow).

Network Center Assistant

• Oct 2017 - Jun 2019

Management and maintenance of school website, networks, and public servers (Ubuntu, CentOS, Java Web).

Graduate Assistant

• Sep 2016 - Jun 2019

Deep Learning Object Detection (Python, MXNet): Designed a lightweight deep-learning-based object detection model for mobile devices. We integrated a convolutional neural network SqueezeNet into the Single-Shot MultiBox Detector.Land Use Machine Learning (Python, Scikit-learn): Implemented an ensemble learning model by stacking machine learning algorithms such as Support Vector Machine and Random Forest to predict land use expansion. The model is implemented using Scikit-learn in Python.Big Geo-Data Processing (Python/Qt): Developed a geospatial big data processing system using Qt, Python and open-source libraries such as Numpy, GDAL/OGR, and Shapely. Algorithms are parallelized and accelerated using multiprocessing and Cython.Outdoor Augmented Reality (Java): Proposed an outdoor augmented reality method combining objection detection, inertial/magnetic sensors, GPS, and spatial relationships. Developed an Android prototype in Java. Show less

Undergraduate Researcher

• May 2014 - Jun 2015

Road Geohazard Early Warning (C++/Qt, GDAL/OGR): Developed a road geohazard early warning system. Implemented several algorithms to extract the geohazard features (e.g. terrain slope and aspect) from large-scale satellite images. SuperMap Cup National University GIS Contest: Evaluation of Wuhan Metro Lines Based on GIS (The 3rd Prize, National; The 1st Prize, Wuhan) Road Geohazard Early Warning (C++/Qt, GDAL/OGR): Developed a road geohazard early warning system. Implemented several algorithms to extract the geohazard features (e.g. terrain slope and aspect) from large-scale satellite images. SuperMap Cup National University GIS Contest: Evaluation of Wuhan Metro Lines Based on GIS (The 3rd Prize, National; The 1st Prize, Wuhan)