Semi-Supervised Learning - Inference

My presentation focused on a general framework to ensure robustness and efficiency in incorporating unlabeled data into statistical models. This is based on two recent research; “Prediction-powered inference(2023)” by Angelopoulos, A. N., Bates, S., Fannjiang, C., Jordan, M. I., & Zrnic, T. and “Doubly-Robust Self-Training(2024)” by Zhu, B., Ding, M., Jacobson, P., Wu, M., Zhan, W., Jordan, M., & Jiao, J.

Basic Research Laboratory Seminar