![[2602.13128] Eventizing Traditionally Opaque Binary Neural Networks as 1-safe Petri net Models](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.13128] Eventizing Traditionally Opaque Binary Neural Networks as 1-safe Petri net Models
Summary
This article presents a framework for enhancing the transparency of Binary Neural Networks (BNNs) by modeling their operations as event-driven processes using Petri nets, facilitating formal verification and analysis.
Why It Matters
As BNNs gain traction for their efficiency in machine learning applications, understanding their opaque behaviors is crucial, especially in safety-critical areas. This framework allows for better verification and transparency, addressing concerns about reliability and accountability in AI systems.
Key Takeaways
- Introduces a Petri net framework to model BNN operations.
- Enhances causal transparency and enables formal verification of BNNs.
- Addresses the challenges of non-linearity and opacity in BNNs.
- Validates the model against software-based BNNs for reliability.
- Supports scalability and complexity assessment in AI systems.
Computer Science > Machine Learning arXiv:2602.13128 (cs) [Submitted on 13 Feb 2026] Title:Eventizing Traditionally Opaque Binary Neural Networks as 1-safe Petri net Models Authors:Mohamed Tarraf, Alex Chan, Alex Yakovlev, Rishad Shafik View a PDF of the paper titled Eventizing Traditionally Opaque Binary Neural Networks as 1-safe Petri net Models, by Mohamed Tarraf and 2 other authors View PDF HTML (experimental) Abstract:Binary Neural Networks (BNNs) offer a low-complexity and energy-efficient alternative to traditional full-precision neural networks by constraining their weights and activations to binary values. However, their discrete, highly non-linear behavior makes them difficult to explain, validate and formally verify. As a result, BNNs remain largely opaque, limiting their suitability in safety-critical domains, where causal transparency and behavioral guarantees are essential. In this work, we introduce a Petri net (PN)-based framework that captures the BNN's internal operations as event-driven processes. By "eventizing" their operations, we expose their causal relationships and dependencies for a fine-grained analysis of concurrency, ordering, and state evolution. Here, we construct modular PN blueprints for core BNN components including activation, gradient computation and weight updates, and compose them into a complete system-level model. We then validate the composed PN against a reference software-based BNN, verify it against reachability and structural ch...
