De-Linearizing Agent Traces: Bayesian Inference of Latent Partial Orders for Efficient Execution

TL;DR

This paper presents BPOP, a Bayesian method for inferring latent dependency structures from noisy agent traces, enabling more efficient execution by pruning irrelevant context and reducing resource usage.

Contribution

BPOP introduces a novel Bayesian inference framework that models traces as stochastic linear extensions, improving dependency recovery and execution efficiency in procedural workflows.

Findings

BPOP outperforms trace-only and process-mining baselines in dependency recovery.

Inferred graphs enable pruning irrelevant context, reducing token usage.

Significant reductions in execution time observed on cloud provisioning and scientific workflows.

Abstract

AI agents increasingly execute procedural workflows as sequential action traces, which obscures latent concurrency and induces repeated step-by-step reasoning. We introduce BPOP, a Bayesianframework that infers a latent dependency partial order from noisy linearized traces. BPOP models traces as stochastic linear extensions of an underlying graph and performs efficient MCMC inference via a tractable frontier-softmax likelihood that avoids #P-hard marginalization over linear extensions. We evaluate on our open-sourced Cloud-IaC-6, a suite of cloud provisioning tasks with heterogeneous LLM-generated traces, and WFCommons scientific workflows. BPOP recover dependency structure more accurately than trace-only and process-mining baselines, and the inferred graphs support a compiled executor that prunes irrelevant context, yielding substantial reductions in token usage and execution time.