Incomplete Information
Incomplete information poses a significant challenge across diverse fields, from artificial intelligence and decision-making to market analysis and scientific modeling. Current research focuses on developing methods to address this challenge, employing techniques such as agent-based systems, Monte Carlo tree search, generative adversarial networks, and dynamic programming to make predictions and decisions despite missing data or uncertain knowledge. These advancements are improving the robustness and efficiency of algorithms in various applications, including autonomous coordination, question answering systems, and optimization problems under uncertainty. The resulting methodologies are enhancing our ability to analyze complex systems and make informed decisions even when complete information is unavailable.
Papers
Federated Learning for Discrete Optimal Transport with Large Population under Incomplete Information
Navpreet Kaur, Juntao Chen, Yingdong Lu
Quantifying Knowledge Distillation Using Partial Information Decomposition
Pasan Dissanayake, Faisal Hamman, Barproda Halder, Ilia Sucholutsky, Qiuyi Zhang, Sanghamitra Dutta