Projects
2025
GPT-OSS for Robot Arm Control
Lu, Y., Manish, S.
I switched the original GPT-4o model into GPT-OSS open source model, and tested
the performance of robot arm control task. Specifically, according to human language,
the LLM is expected to generate specific python code to control the robot arm, to perform
certain tasks.
Multi-task Hydro-Net model
Rohan, R., Lu, Y., Yong, Y., Le, M.
derive the object detection, depth estimation, and semantic segmentation at the same time
2024
LLM GPT-2 fine tune experiment
Lu, Y., Li, X.
It is a project for fine-tuning GPT-2 model on custom text dataset. The main goal is to test the performance change
of GPT-2 after fine-tuning on a small-scale dataset, using Python and the Hugging Face Transformers library.
2023
Horse Racing Forecast
Lu, Y.
Given the horse racing data from previous years, several time series models were used to test
the performance of the next year.
RLHF for route design problem
Lu, Y., Wang, Y., Xie, J., Ye, F.
We adjusted Q-learning algorithm with human feedback, and tested it in route design problem.
NBA data processing
Lu, Y., Shi, W., Liu, Y.
We wrote a R package to process NBA data. We helped visualize and analyze players' performance.
Image Denoising Filter Design
Lu, Y., Li, J.,
It is a project designing image denoising filters with innovative ideas on NASA images.