MTI-GNN Models Cognitive Triad Interactions for Distortion Detection

MTI-GNN detects cognitive distortions by modeling Beck's triad—self, world, and future—using LLM-driven perspective decomposition and Graph Neural Networks. The architecture uses a Triad Interaction module to capture cross-perspective dependencies for improved mental health classification.