Resources

Enabling the world's first at-home recycling solution.

June 2026

Improved edge AI performance for autonomous robotics workloads.

April 2026

A case study in smarter neural architectures for enterprise AI systems.

March 2026

Demonstrating dendritic intelligence in real-world edge AI settings.

December 2025

Introducing the third dimension in neural network architecture for model efficiency.

Exploring how pruning and dendritic perforation work together to create smaller, more efficient neural networks.

How Perforation Lets ResNet-18 Perform Like ResNet-34

Making powerful AI models smaller, faster, and deployable on constrained edge devices.

Researchers found that brain-inspired dendritic structures can reduce parameter count while improving accuracy.

Real-time human detection for industrial robotics, with fewer parameters and faster deployment.

A biologically inspired extension to backpropagation that improves performance with less compute.

How brain-inspired dendritic features are making artificial neural networks more powerful.

Novel approaches to classification, approximation, and prediction using dendritic models.

Foundational work on Cascade-Correlation as an architecture and supervised learning algorithm for artificial neural networks.

An introduction to how dendritic perforation adds neuron-specific learning signals to neural networks, enabling higher accuracy with fewer parameters and less data.

Exploring dendritic learning and efficient neural networks with Qualcomm and Edge Impulse.

Exploring real-time AI, sports analytics, and next-generation machine learning during NFL Draft week.

Perforated is implemented entirely in Python using standard PyTorch functionality.

Conversation with Thought Liters about the future of AI.

Why the next generation of machine learning may depend on smarter neural architectures, not just larger models.