Dong WANG

Hi👋 I am an undergraduate student pursuing a Bachelor’s Degree in Computer Science at the College of Science, Purdue University, West Lafayette.

I led a project titled “Video Prediction through Physical Laws” under the guidance of Ph.D. candidate Max and Assistant Professor Yexiang Xue during the summer 2024. Existing models for prediction often need large datasets and struggle with complex motions, especially with limited data. We present a framework for predicting object motion by leveraging known physical laws to reduce the reliance on large datasets and enhances the model’s ability to predict complex and abrupt motions.

I joined the Vertically Integrated Program from January 2024 to May 2024, where I worked with the Purdue Aerial Robotics Team under the IEEE chapter’s PART committee. I introduced hand-drawn sketches as a more precise input method for drone-based item transport, trained a GAN-based neural network for image-to-sketch conversion.

Before transferring to Purdue University, I studied Communication Engineering at Northeastern University (China) for one year and a half. There, I had the honor of serving as a Teaching Assistant under the guidance of Dr. Peng Han in Fall 2023.

🌱 So far, I have attempted some interesting personal projects, such as:

• EchoGen: An iterative optimization system that refines LLM-generated results through iterative feedback.
• SmartGA: Using LLM to guide the direction of the evolution of the Genetic Algorithm.
• SketchTune: Fine-tuning the CLIP model on the sketch (The Sketchy Database).
• Gmaillm: Utilizing LLM to summarize important Gmail messages and discard irrelevant ones.