Pairwise Influence

Pairwise influence examines how individual elements within a system affect each other, impacting overall system behavior. Current research focuses on understanding and mitigating these influences across diverse applications, including improving the accuracy and efficiency of machine learning models (e.g., through optimized weight sharing in continual learning and addressing serial position effects in large language models), enhancing the speed and accuracy of algorithms (e.g., in Boolean function classification), and analyzing information flow in social networks and dynamic environments. This research is significant for improving the performance and robustness of various systems, from AI models to social network analysis and robotic perception.