Neuro-symbolic control and stochastic Petri nets
The modelling substrate that runs through the whole federation: learned components inside a symbolic, analysable net.
A Petri net is a formal model of concurrency: places hold tokens, transitions move them, and the reachable markings describe every state the system can enter. Stochastic Petri nets attach firing rates to transitions, turning the net into a continuous-time Markov process whose performance and safety properties can be analysed rather than only simulated. Generalised variants mix instantaneous logical transitions with timed stochastic ones.
Neuro-symbolic control pairs this analysable symbolic backbone with learned components. The symbolic net carries the constraints and the guarantees; the learned parts supply adaptation and estimation where a closed-form model is unavailable. The result is a controller whose behaviour can still be reasoned about, because the symbolic layer bounds what the learned layer is allowed to do.
This pairing is the signature of the federation. The plasma controllers, the fusion suites, and the merge fabric are all described as neuro-symbolic stochastic Petri-net systems — which is why an evidence contract that shows a claim at its honest boundary matters here: the symbolic net is where a bounded guarantee is actually stated.
The vocabulary of the topic.
- Place / transition / token
- The three primitives of a Petri net; markings of tokens are the system's states.
- Stochastic firing rate
- A rate attached to a transition, making the net a continuous-time Markov process.
- Reachability
- The set of markings a net can enter; the basis of its safety and liveness analysis.
- Neuro-symbolic pairing
- A learned component constrained by an analysable symbolic model that bounds its behaviour.
Studios working on this.
Each runs standalone in its own repository and federates its evidence through the platform.
Run it, don't take it on faith.
The runnable lessons and sealed measurements this site ships for the topic.
Read it at source, in order.
Canonical references for the topic, ordered from the foundations to current work. Every one was verified at source — a DOI resolves through doi.org, a standard through its issuer — so each link goes to the real record.
Start with the foundations
- 1
The paper that attached exponential firing rates to Petri nets, making them a stochastic model.
How it works
- 2
Generalised stochastic Petri nets (GSPN) — immediate plus timed transitions.
- 3
The standard tutorial on Petri-net structure, reachability, and invariants.
Current work
- 4
The case for combining learned and symbolic components — the pattern the controllers follow.
Related topics.
Where this topic connects to the rest of the federation's work.