PRIM: Meta-Learned Bayesian Root Cause Analysis

TL;DR

PRIM is a meta-learned Bayesian approach for root cause analysis that efficiently identifies causal changes in complex systems without explicit statistical testing, enabling zero-shot inference.

Contribution

It introduces PRIM, a novel causal meta-learning method that implicitly learns causal structure and performs zero-shot root cause analysis in complex systems.

Findings

PRIM achieves zero-shot inference in 17 ms for systems with up to 100 variables.

PRIM outperforms graph-unaware methods on synthetic and real benchmark datasets.

Lightweight fine-tuning further improves PRIM's performance.

Abstract

Root cause analysis (RCA) in complex systems is challenging due to error propagation across multiple variables, the need for structural causal knowledge, and the computational cost of inference at test time. We introduce PRIM (Prior-fitted Root cause Identification with Meta-learning), a causal meta-learning approach that frames RCA as a Bayesian inference task over a synthetic prior of causal models. By marginalising out structural uncertainty, PRIM implicitly identifies changes in the data-generating mechanism between baseline and anomalous periods. In doing so, PRIM infers distributional differences without explicit statistical testing, and implicitly learns causal structure without model fitting at test time. Following the simulation-based meta-learning paradigm of prior-fitted networks, PRIM uses a Model-Averaged Causal Estimation (MACE) transformer neural process that jointly attends over observational and anomalous samples and the causal structure of nodes, enabling zero-shot inference in 17,ms for systems with up to 100 variables. Across synthetic benchmarks and two realistic benchmark datasets, PetShop and CausRCA, PRIM is competitive with methods that are aware of the system's causal graphical structure a priori while outperforming graph-unaware methods on several tasks. Lightweight fine-tuning to specific domains and data dynamics improves performance further.