North System
"System," in various contexts, refers to a structured approach to solving complex problems, ranging from speech recognition to machine learning model explainability. Current research focuses on improving system performance through techniques like diverse outlier sampling for robust out-of-distribution detection and incorporating domain knowledge to enhance safety and efficiency in reinforcement learning. These advancements are significant for improving the reliability and interpretability of AI systems across diverse applications, from natural language processing to safety-critical domains.
Papers
October 5, 2023
June 3, 2023
May 16, 2023
May 10, 2023
April 23, 2023
April 15, 2023