Core Concepts

Deep dive into library design, architecture, and theoretical foundations.

This section provides in-depth explanations of the CANNs library’s design principles, module organization, and conceptual foundations. These documents focus on the “why” and “when” rather than the “how”, helping you understand the library’s architecture and make informed decisions about using its components.

Topics:

• Design Philosophy & Architecture Overview
• Model Collections
• Task Generators
• Analysis Methods
• Brain-Inspired Training

Overview

Design Philosophy & Architecture Overview

Understanding the library’s architecture, core design principles, and the four core application scenarios. Learn about separation of concerns, extensibility, BrainPy integration, and performance strategies.

Model Collections

Explore the three model categories: Basic CANN models, Brain-Inspired models with learning mechanisms, and Hybrid models combining CANN with ANNs. Understand the BrainPy foundation and how to implement custom models.

Task Generators

Task generation philosophy and available paradigms. Learn about tracking tasks (population coding, template matching, smooth tracking) and navigation tasks (closed-loop, open-loop). Understand model-task coupling and design considerations.

Analysis Methods

Comprehensive analysis tools including Model Analyzer for simulations, Data Analyzer for experimental recordings, RNN Dynamics Analysis for fixed points, and Topological Data Analysis for geometric structures.

Brain-Inspired Training

Brain-inspired learning mechanisms and the Trainer framework. Understand activity-dependent plasticity, learning rules (Hebbian, STDP, BCM), and how to implement custom trainers for biologically plausible learning.