# 🔬 Foundations

A good AI engineer or researcher must understand the building blocks. Here's a structured list of core and cutting-edge topics:

## ✅ Core Concepts

* **Perceptron, ReLU**\
  How a network shares data and processes it.
* **Gradient Descent, Adam Optimizer, Backpropagation**\
  The equations of learning.
* **Loss Functions**\
  MSE, Cross-Entropy, KL Divergence, etc.
* **Feedforward Neural Networks, CNNs, RNNs, etc**\
  The most classical concepts in deep learning.

## ✅ Advanced Topics

* **Variational Autoencoders (VAE)**\
  A way to compress information in a network and analyze how the AI interprets data. Foundation for latent space understanding and Nomic AI (the beginning of embeddings).
* **Generative Adversarial Networks (GANs)**\
  The real beginning of AI in art. Still used in modern architectures like diffusion models and various loss strategies.
* **Contrastive Language–Image Pre-training (CLIP)**\
  Foundation of multimodal learning.
* **Large Language Models (LLMs)**\
  GPT, LLaMA, Qwen, etc.
* **Mixture of Agents**\
  Example: DeepSeek. Combining multiple models to act as one and select the best outputs.
* **Multimodal Systems**\
  Gato, Gemini, Flamingo, etc.

### Reinforcement Learning (RL)

* **Deep Q-Learning**
* **Policy Gradients**
* **PPO, A3C, SAC**, etc.

### Other Concepts

* **Latent Spaces**\
  Interpretation & manipulation.
* **Prompt Engineering & Evaluation**
* **Verification Models**\
  Output confidence & self-verification.
* **XAI (Explainable AI)**\
  Making models understandable and proving they are unbiased.
* **Graph Neural Networks (GNNs)**\
  An innovative approach used by Google. Less feedforward, more interconnected.
* **Liquid Neural Networks**\
  Currently more theoretical. Hard to understand and train, with minor gains. Considered inefficient by many, but intriguingly beautiful.


---

# Agent Instructions: 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:

```
GET https://doc.guidry-cloud.com/00-foundation.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
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.