IMAGINE: Metacognitive Counterfactual Reasoning for Robust Adversary System Analysis
Proceedings of the AAAI Symposium Series · 06.11.2026 · 10 (1) · 75-83
DOI: 10.1609/aaaiss.v10i1.43393 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Symposium Series, online erschienen am 06.11.2026. Verfasst von Jaehoon Choe und 3 weiteren Autor:innen.
Abstract (Originalsprache)
The development of more robust and adaptable decision support systems is crucial for handling real-world applications where data is often noisy, uncertain, or dynamic. One of the key challenges in developing such systems is enabling them to adapt to variations in data, environment, or tasks. This paper proposes a novel approach, IMAGINE, which leverages metacognitive principles and contrastive learning techniques to enhance the cognitive capabilities of existing decision support systems. We demonstrate the effectiveness of IMAGINE in the context of Adversarial System Analysis (ASA), and show how it can improve the performance of decision support systems in handling variability. Our approach has broader implications for improving decision support systems across various applications, and we discuss future research directions such as quantifying the effects of coordinates, or interactive decomposition.
Zitieren
APA
Choe, J., Guo, J. K., Hubchak, M., & Lu, T. (2026). IMAGINE: Metacognitive Counterfactual Reasoning for Robust Adversary System Analysis. Proceedings of the AAAI Symposium Series, 10(1), 75-83. https://doi.org/10.1609/aaaiss.v10i1.43393
Vancouver
Choe J, Guo JK, Hubchak M, Lu T. IMAGINE: Metacognitive Counterfactual Reasoning for Robust Adversary System Analysis. Proceedings of the AAAI Symposium Series. 2026;10(1):75-83. doi:10.1609/aaaiss.v10i1.43393
BibTeX
@article{Choe2026,
author = {Choe, Jaehoon and Guo, Jinhong K. and Hubchak, Mykola and Lu, Tsai-Ching},
title = {IMAGINE: Metacognitive Counterfactual Reasoning for Robust Adversary System Analysis},
journal = {Proceedings of the AAAI Symposium Series},
year = {2026},
volume = {10},
number = {1},
pages = {75--83},
doi = {10.1609/aaaiss.v10i1.43393}
}
EUVIMED