Neural Computer Architecture

This project represents a groundbreaking collaboration between the University of Zagreb and Beijing Normal University, focusing on implementing basic cognitive mechanisms in technically feasible systems. Our goal is to create computer architectures that are directly inspired by biological neural networks.

Project Overview

Traditional computer architectures are based on the von Neumann model, which separates processing and memory. However, biological neural networks operate on fundamentally different principles that offer unique computational advantages.

Key Biological Inspirations

Synaptic Plasticity

• Connections between neurons can strengthen or weaken based on activity patterns
• Enables learning and adaptation without external programming
• Forms the basis for associative memory and pattern recognition

Content-Addressable Memory

• Information retrieval based on content rather than memory addresses
• Allows for robust pattern completion from partial cues
• Enables fault-tolerant operation even with damaged components

Technical Implementation

Hardware Considerations

Our implementation focuses on creating systems that maintain biological plausibility while achieving practical computational performance:

• Distributed processing units that combine computation and memory
• Adaptive connection weights that modify based on usage patterns
• Parallel architecture that enables massive concurrent processing

Software Framework

We’re developing a software framework that enables:

• Real-time synaptic weight modification
• Pattern-based memory retrieval
• Emergent computational behaviors

Research Applications

This architecture has potential applications in:

Neuromorphic Computing

Creating brain-inspired processors that are more energy-efficient than traditional computers, particularly for pattern recognition and associative memory tasks.

Cognitive Modeling

Developing computational models that can help us understand how biological cognition emerges from neural network dynamics.

Adaptive Systems

Building systems that can learn and adapt to new environments without explicit reprogramming.

Current Progress

• ✅ Theoretical framework established
• ✅ Basic synaptic plasticity models implemented
• 🔄 Content-addressable memory system under development
• 📋 Hardware prototyping planned for late 2025

This work builds upon my paper “Biological Neural Machines – Creating a Computer on Biologically Restrained Neural Networks” presented at the IRCN and Chen Institute Joint Course on Neuro-inspired Computation.