Research
Machine Learning & Computer Vision Researcher | Army Research Laboratory
09/2021 - 05/2023
Research focus: Unsupervised Domain Adaptation, Synthetic-to-real Translation; Professor: Suya You
- Trained and evaluated different setups of a synthetic-to-real generative model that transforms the synthetic renderings into more realistic styles on general-purpose datasets conditioned with unlabeled real-world data
- Used Tensorflow and Keras to implement a Generative Adversarial Network
- Conducted extensive performance evaluations through FID and SSIM metrics and through downstream object detection/semantic segmentation tasks
Undergraduate Researcher | Berkeley Lab
09/2021 - 12/2022
Research focus: Scientific Machine Learning(SciML), Surrogate Modeling; Professor: Duncan Callaway
- Explored the application of continuous-time echo state networks (CTESNs) for learning a surrogate to predict the time-domain solution of a dynamic power system for varying system parameters
- Implemented non-linear projection CTESN using Julia and benchmark it against the linear projection CTESN
- Implemented cross-validation based adaptive sampling and compare it with gradient-based and query-by-committee active sampling; benchmarked their accuracy and computational costs against a simple space-filling approach
- Evaluated the performance on HPC for a dynamic power system application
Undergraduate Researcher | Sutardja Center for Entrepreneurship and Technology
01/2021 - 12/2021
Research job: Implement and Maintain SCET’s innovation index survey system; Professor: Ikhlaq Sidhu
- Used webhooks and Google app scripts to create a survey system that collects, processes surveys, and gives feedback to the users; built an efficient pipeline that could handle concurrent requests
- Deployed the survey on Anvil; built an interactive interface with a recommender system for survey distributors
- Customized the system to other entities
Undergraduate Researcher | Creative Commons
09/2020 - 06/2021
Research topic: image popularity and authority; Github: repo; Mentor: Brent Moran
- Used Python, SQL, pandas to scrape, process, and organize PBs of image data
- Embedded image data into vectors using Bert Model and ran MLP Regressor on them to train and predict
- Extracted a single layer from VGG16 model to recognize and classify large image data