Wanjing Ma

Navigating Data Corruption in Machine Learning: Balancing Quality, Quantity, and Imputation Strategies

Qi Liu, Wanjing Ma

Data corruption, including missing and noisy entries, is a common challenge in real-world machine learning. This paper examines its impact and mitigation strategies through two experimental setups: supervised NLP tasks (NLP-SL) and deep reinforcement learning for traffic signal c…Read more