Long Yunbo is a Ph.D. student in the Department of Engineering at the University of Cambridge, specializing in the generation of tabular and graph data. His research addresses both the scaling up and compression of data. For scaling up, he explores synthesizing heterogeneous and time-series data using autoencoders and deep diffusion generative models to create new datasets based on original data. He also study how to leverage large language models, such as GPT2 or Llama 3, for fine-tuning to generate custom datasets by better understanding text content and developing a universal data generation model to streamline data processing. For data compression, Yunbo focuses on data distillation in hyperbolic space to reduce redundancy in large-scale datasets and optimize data for specific machine learning tasks.

Before this, Yunbo completed a Master’s degree in Engineering Department at the University of Cambridge, where his research focused on a differential privacy-based personalized clustering federated learning method. Prior to that, he earned a First-Class Honours degree in Engineering from the University of Birmingham, with his undergraduate thesis exploring the application of deep reinforcement learning algorithms for autonomous robotic dismantling tasks.