> For the complete documentation index, see [llms.txt](https://neurosymbolicai.gitbook.io/neuro-symbolic-ai-in-practice/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://neurosymbolicai.gitbook.io/neuro-symbolic-ai-in-practice/part-iii-core-approaches/chapter-4/4-3-hybrid-architectures.md).

# 4.3 Hybrid / Co-Processing Architectures

> *"These setups treat both neural and symbolic systems as co-equal components, running them in tandem or passing structured data back and forth to leverage the best of both worlds."*

***

Hybrid co-processing architectures are the most powerful and most demanding neuro-symbolic pattern — and the fastest-growing area of NeSy research.(Anon., 2025a) Unlike §4.1 (where neural is primary) and §4.2 (where symbolic is primary), hybrid architectures are fundamentally **bidirectional**: information flows from neural to symbolic *and* from symbolic to neural in a tight, continuous loop.

The key design principles, established across all systems in this section:

1. **Define the interface precisely.** The neural-to-symbolic boundary must be a well-specified data structure — a scene graph, a PDDL file, a proof context, a KG triple. Ambiguity at the interface is the most common source of failure.
2. **Let the symbolic component verify.** The neural component proposes; the symbolic component verifies. Never let neural outputs flow unchecked to downstream actions.
3. **Use symbolic feedback for neural improvement.** Verified proofs, correct plans, and consistent KG triples become training signal — no human labeling required.
4. **Co-design the world model with the planner.** Model accuracy in the wrong regions of state space does not help. Design the world model to be accurate where the planner needs it.

***

## Sections in This Chapter

* [4.3.1 System 1 / System 2 Integration — AlphaProof and AlphaGeometry](/neuro-symbolic-ai-in-practice/part-iii-core-approaches/chapter-4/4-3-hybrid-architectures/4-3-system1-system2.md)
* [4.3.2 Knowledge Graph Integration](/neuro-symbolic-ai-in-practice/part-iii-core-approaches/chapter-4/4-3-hybrid-architectures/4-3-knowledge-graphs.md)
* [4.3.3–4.3.8 Case Studies and Applications](/neuro-symbolic-ai-in-practice/part-iii-core-approaches/chapter-4/4-3-hybrid-architectures/4-3-applications.md) — AlphaCode, PLOI, FunSearch, PDDLStream, DreamerV3, and more

***

## References

1. Anon. "Neurosymbolic AI: A Survey." *arXiv preprint* arXiv:2508.13678 (2025a). <https://arxiv.org/abs/2508.13678>


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://neurosymbolicai.gitbook.io/neuro-symbolic-ai-in-practice/part-iii-core-approaches/chapter-4/4-3-hybrid-architectures.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.