Metacognitive Composable AI (MCAI)

MCAI was funded as a DARPA Collaborative agreement from 2024-2025 led by Paulo Shakarian. Below are listed papers related to the key findings of that work.

Preprints

M. Leiva, N. Ngu, J. S. Kricheli, A. Taparia, R. Senanayake, P. Shakarian, N.D. Bastian, J. Cororan, G. Simari, Consistency-based Abductive Reasoning over Perceptual Errors of Multiple Pre-trained Models in Novel Environments.

A. Taparia, N. Ngu, M. Leiva, J. S. Kricheli, J. Cororan, N.D. Bastian, G. Simari, P. Shakarian, R. Senanayake, VLC Fusion: Vision-Language Conditioned Sensor Fusion for Robust Object Detection.

Published / Accepted

N. Ngu, A. Taparia, G. Simari, M. Leiva, J. Corcoran, R. Senanayake, P. Shakarian, N. Bastian, Multiple Distribution Shift – Aerial (MDS-A): A Dataset for Test-Time Error Detection and Model Adaptation, AAAI Spring Symposium (AAAI-MAKE) (Apr. 2025).

P. Shakarian, G. Simari, N. Bastian, Probabilistic Foundations for Metacognition via Hybrid-AI, AAAI Spring Symposium (AAAI-MAKE) (Apr. 2025).

D. Aditya, C. Payne, M. A. Leiva, P. Shakarian, Machine Learning Model Integration with Open World Temporal Logic for Process Automation, 41st Intl. Conference on Logic Programming (ICLP-25).