Multi-Level Causal Embeddings

2026-02-25

AI Summary

This work addresses the challenge of unifying multiple fine-grained causal models with heterogeneous representations into a single coarse-grained causal model while preserving causal consistency. To this end, the authors propose a causal embedding framework—a generalized formulation of causal abstraction—that formally characterizes the mechanism for integrating multi-level causal models. The framework introduces the multi-resolution marginal problem to jointly reconcile data integration at both statistical and causal levels. By synthesizing principles from causal modeling, abstraction theory, and multi-resolution analysis, the approach provides theoretical guarantees for the fidelity of cross-granularity causal structures and enables effective fusion of heterogeneous model data.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Web Mining and Content Analysis: Bridging structured and unstructured data
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Abstractions of causal models allow for the coarsening of models such that relations of cause and effect are preserved. Whereas abstractions focus on the relation between two models, in this paper we study a framework for causal embeddings which enable multiple detailed models to be mapped into sub-systems of a coarser causal model. We define causal embeddings as a generalization of abstraction, and present a generalized notion of consistency. By defining a multi-resolution marginal problem, we showcase the relevance of causal embeddings for both the statistical marginal problem and the causal marginal problem; furthermore, we illustrate its practical use in merging datasets coming from models with different representations.

Problem

• causal embeddings
• causal models
• abstraction
• marginal problem
• multi-resolution

Innovation

• causal embeddings
• causal abstraction
• multi-resolution marginal problem
• causal consistency
• model coarsening