Pawsterior: Variational Flow Matching for Structured Simulation-Based Inference

2026-02-14

AI Summary

This work addresses the limitations of conventional flow matching methods in simulation-based inference, which struggle with structured constraints—such as bounded parameters and mixed discrete-continuous variables—and cannot model discrete latent variables. The authors propose a variational flow matching framework that explicitly incorporates the geometric structure of the domain into the inference process through endpoint variational modeling. They introduce, for the first time, a formal mechanism of endpoint-induced affine geometric constraints that accommodates discrete latent structures. By integrating geometric inductive biases with an extension of CatFlow, the method achieves substantially improved posterior fidelity on standard SBI benchmarks, demonstrates superior performance in classifier two-sample tests, and successfully handles inference tasks involving discrete structures that are intractable for traditional approaches.

Technology Category

• Constraint Satisfaction and Optimization: Mixed Discrete/Continuous Optimization
• Search and Optimization: Mixed Discrete/Continuous Search
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

We introduce Pawsterior, a variational flow-matching framework for improved and extended simulation-based inference (SBI). Many SBI problems involve posteriors constrained by structured domains, such as bounded physical parameters or hybrid discrete-continuous variables, yet standard flow-matching methods typically operate in unconstrained spaces. This mismatch leads to inefficient learning and difficulty respecting physical constraints. Our contributions are twofold. First, generalizing the geometric inductive bias of CatFlow, we formalize endpoint-induced affine geometric confinement, a principle that incorporates domain geometry directly into the inference process via a two-sided variational model. This formulation improves numerical stability during sampling and leads to consistently better posterior fidelity, as demonstrated by improved classifier two-sample test performance across standard SBI benchmarks. Second, and more importantly, our variational parameterization enables SBI tasks involving discrete latent structure (e.g., switching systems) that are fundamentally incompatible with conventional flow-matching approaches. By addressing both geometric constraints and discrete latent structure, Pawsterior extends flow-matching to a broader class of structured SBI problems that were previously inaccessible.

Problem

• simulation-based inference
• structured posterior
• flow matching
• discrete latent structure
• geometric constraints

Innovation

• variational flow matching
• structured simulation-based inference
• geometric confinement
• discrete latent structure
• endpoint-induced affine constraint