> For the complete documentation index, see [llms.txt](https://alliance-games.gitbook.io/alliance-games-gitbook/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://alliance-games.gitbook.io/alliance-games-gitbook/technical-structure/hybrid-ai-architecture.md).

# Hybrid AI Architecture

Alliance Games’ hybrid AI architecture solution combines local edge processing with cloud-based AI to deliver dynamic, real-time gameplay. This dual-layered approach enables immediate responses to player actions while leveraging cloud resources for complex, data-intensive processing.

***

### Local Edge AI

Edge Nodes and Micro Nodes manage on-device AI processing, allowing for low-latency interactions and personalized gameplay experiences.

* **Advantages**: Delivers responsive, low-latency AI that adjusts to the player’s behavior, creating an immersive experience that feels unique to each player.
* **Sample Application**: In a survival game, NeuralForge adjusts NPC actions based on player decisions, making NPCs feel adaptive and realistic.

***

### Cloud-Based AI

Cloud-based AI handles large-scale processing, including model training and behavioral analysis, to refine AI behaviors and provide consistent experiences.

* **Advantages**: Scalable processing for model refinement and behavioral analysis, enhancing model accuracy and providing adaptive experiences across players.
* **Sample Application**: In an open-world game, cloud-based AI analyzes player trends to generate new events and scenarios based on popular playstyles.

***

## AI-Driven Intelligence

Alliance Games uses sophisticated machine learning models within both **AG Game AI** and **AG Game Cloud** to enable adaptive, intelligent gameplay and efficient resource management. By leveraging supervised, unsupervised, and reinforcement learning techniques, these models power a range of features from NPC behavior to personalized content recommendations.

* **Supervised Learning**: Used primarily in **GameMind** to develop personalized gameplay by analyzing labeled data on player preferences, skill levels, and historical actions.
* **Unsupervised Learning**: Enables GameMind to discover patterns in player behavior, such as common strategies or preferred challenges, without specific labels.
* **Reinforcement Learning**: Employed by **NeuralForge** to enable NPCs to make context-aware decisions, adapting their behavior based on positive or negative feedback from interactions with players.

Alliance Games provides a responsive AI system that adapts in real-time while leveraging scalable cloud resources to manage computationally intensive tasks.