> 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.md).

# Chapter 4: Neuro-Symbolic AI in Practice

> *"The best of both worlds is not a compromise — it is a synthesis."*

Neuro-symbolic AI combines the pattern-recognition strengths of neural networks with the logical reasoning and explicit rule-handling of symbolic AI. If Chapter 3 built the formal vocabulary, this chapter shows what to do with it: concrete architectures, implementation patterns, and production deployments organized around four complementary roles that neural and symbolic components can play relative to each other.

## The Four Architectural Categories

This chapter organizes the neuro-symbolic design space into four categories (Garcez & Lamb, 2023; Sheth et al., 2023):

1. **Symbolic Helps Neural** — Symbolic rules and knowledge constrain, guide, or structure neural learning. The symbolic component is the guardrail; the neural component is the engine.
2. **Neural Helps Symbolic** — Neural networks accelerate, approximate, or replace computationally expensive symbolic operations. The symbolic framework retains its formal guarantees; the neural component provides speed.
3. **Hybrid / Co-Processing Architectures** — Both systems operate as co-equals, passing structured information between them in a tight loop.
4. **Pure Neural World Models** — Approaches without explicit external symbolic structure: either the abstract domain IS the neural representation (§4.4.1 Differentiable Abstract Interpretation), or planning operates entirely in continuous learned feature space (§4.4.2 JEPA World Models). These systems are the canonical neural-only alternatives that the neuro-symbolic community must engage with.

These categories are not mutually exclusive — the most powerful systems often combine elements of multiple paradigms. But as a taxonomy, they provide a practical design vocabulary for architects building neuro-symbolic systems.

## Navigating This Chapter

* [4.1 Symbolic Helps Neural](/neuro-symbolic-ai-in-practice/part-iii-core-approaches/chapter-4/4-1-symbolic-helps-neural.md)
* [4.2 Neural Helps Symbolic](/neuro-symbolic-ai-in-practice/part-iii-core-approaches/chapter-4/4-2-neural-helps-symbolic.md) — including:
  * [4.2.1 Differentiable Reasoning](/neuro-symbolic-ai-in-practice/part-iii-core-approaches/chapter-4/4-2-neural-helps-symbolic/4-2-differentiable-reasoning.md)
  * [4.2.2 Neural Subroutines in Symbolic Systems](/neuro-symbolic-ai-in-practice/part-iii-core-approaches/chapter-4/4-2-neural-helps-symbolic/4-2-neural-subroutines.md)
  * [4.2.3 Answer Set Programming](/neuro-symbolic-ai-in-practice/part-iii-core-approaches/chapter-4/4-2-neural-helps-symbolic/4-2-asp.md)
* [4.3 Hybrid / Co-Processing Architectures](/neuro-symbolic-ai-in-practice/part-iii-core-approaches/chapter-4/4-3-hybrid-architectures.md)
* [4.4 Pure Neural World Models](/neuro-symbolic-ai-in-practice/part-iii-core-approaches/chapter-4/4-4-pure-neural.md)

***

## References

1. Garcez, Artur d'Avila, and Luís C. Lamb. "Neurosymbolic AI: The 3rd Wave." *Artificial Intelligence Review* 56 (2023): 12387–12406. <https://doi.org/10.1007/s10462-023-10448-w>
2. Sheth, Amit, Kaushik Roy, and Manas Gaur. "Neurosymbolic Artificial Intelligence (Why, What, and How)." *IEEE Intelligent Systems* 38.3 (2023): 56–62. <https://doi.org/10.1109/MIS.2023.3234994>


---

# 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.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.