VoxLogicA2: A Spatial Model Checker for Neuro-symbolic Medical Imaging (extended version)

Gina Belmonte; Laura Bussi; Vincenzo Ciancia; Mieke Massink

Home/Archive/arXiv

arXiv · 2610.26626

VoxLogicA2: A Spatial Model Checker for Neuro-symbolic Medical Imaging (extended version)

Gina Belmonte·Laura Bussi·Vincenzo Ciancia·Mieke Massink

View original source ↗Read source document ↗Export metadata ↓

Abstract

We present the tool VoxLogicA2, a spatial model checker aimed at Computer Vision, with applications in high-resolution 2D and 3D medical imaging. The new version of the tool combines logical primitives, non-logical primitives, and training and prediction primitives of state-of-the-art neural networks, freely interleaved with the symbolic tool functionality. The primary application field is neuro-symbolic medical image analysis. VoxLogicA2 is a fast, parallelising spatial model checker with a declarative domain-specific language, capable of describing both the learning procedures and the spatial logic procedures of an analysis in one program. We validate our approach by verifying its improved efficiency, and by using the new constructs of the language to describe case studies of previous work, which were based on Python, with a specification that uses VoxLogicA2 only

Explore related subjects

cs.LO

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Gina Belmonte, Laura Bussi, Vincenzo Ciancia, Mieke Massink. 2026-10-19. VoxLogicA2: A Spatial Model Checker for Neuro-symbolic Medical Imaging (extended version). https://arxiv.org/abs/2610.26626

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Forall-Exists Relational Verification by Filtering to Forall-Forall

Relational verification encompasses reasoning about data abstraction, security and privacy, secure compilation, and functional specification of tensor programs, among others. Several relational logics and tools support compositional reasoning about $\forall\forall$ (2-safety) properties and, generally, $k$-safety properties. In contrast, few logics and tools exist for reasoning about $\forall\exists$ properties which are critical in the presence of nondeterminism. This paper's primary contribution is a technique for verifying a $\forall\exists$ judgment by way of a novel transformation. The filter-adequacy transformation adds assertions to a product program in such a way that the desired $\forall\exists$ property (of a pair of underlying unary programs) is implied by a $\forall\forall$ property of the transformed product. The paper develops bicoms, a form of product programs that facilitates expressing good alignments and caters for direct translation of $\forall\forall$ properties to unary correctness; proves that successful $\forall\forall$ verification of a transformed bicom implies the $\forall\exists$ spec for its underlying unary commands; and evaluates the technique using a prototype auto-active relational verifier. The technique enables users to work with ordinary assertions and assumptions, and a standard assertion language, so that existing $\forall\forall$ and unary verification tools can be used for $\forall\exists$.

cs.LO↗

arXiv2026

Synthetic Differential Geometry in Lean

This article is about the formalization of synthetic differential geometry with the Lean proof assistant and the mathematical library mathlib. The main result we prove and formalize is a Taylor theorem for functions of several variables, where the series expansion is around an infinitesimal neighborhood. Most of our proofs are in fact new. Our investigations highlight the possibility of using mathlib to do constructive mathematics.

cs.LO↗

arXiv2026

Reasoning about Continuous-Variable Quantum Systems

Continuous-variable quantum computing (CVQC) is a computing paradigm in which measurements yield values over a continuous domain. CVQC is both a convenient computational framework for modeling physical quantum systems, and a good abstraction for hardware platforms based on quantum optics. Yet, the semantic foundations of CVQC remain underdeveloped. To address this gap, we develop a formal semantics for a core CV quantum programming language, and sound verification methods for program correctness. A main contribution of this work is to isolate a well-behaved quantitative predicate domain that achieves sufficient expressiveness to accommodate unbounded values as they arise in the infinite-dimensional, continuous setting. Specifically, we choose closed positive quadratic forms as semantic predicates, representing finite expectations, domains of finiteness, and infinite penalties in one ordered object. We validate our choice by establishing that our semantic predicates satisfy desirable closure properties including the definition of weakest preconditions. We validate our design with two case studies, including an example based on the celebrated GKP error-correcting code, for which we establish a second moment bound.

cs.LO↗