Relation-Oriented: Toward Knowledge-Aligned Causal AI

This study examines the inherent limitations of the prevailing Observation-Oriented learning paradigm by understanding relationship modeling from a unique dimensionality perspective. This paradigm necessitates the identification of modeling objects prior to defining relations, confining models to observational space, and limiting their access to dynamical temporal features. By relying on a singular, absolute timeline, it often neglects the multi-dimensional nature of the temporal feature space. This oversight compromises model robustness and generalizability, contributing significantly to the AI misalignment issue. Drawing from the relation-centric essence of human cognition, this study presents a new Relation-Oriented paradigm, complemented by its methodological counterpart, the relation-defined representation learning, supported by extensive efficacy experiments.