Neuromorphic computing

Lacuna

Active development

A general-purpose spiking neural network simulation framework for efficient neuromorphic computing and flexible biological modeling.

Overview

Lacuna provides the simulation foundation for the lab’s work on spiking neural networks and neuromorphic computing. It represents networks as arbitrary directed graphs with heterogeneous, equation-defined neuron and synapse models. A capability-driven compiler analyzes those models and selects execution strategies based on their mathematical structure.

Capabilities

Event-driven analytical propagation, numerical threshold-crossing search, stepped numerical integration, and mixed-mode execution within one network
High-level Python model authoring and a native C interface backed by a portable C runtime
Execution on conventional computing systems and resource-constrained microcontrollers
Integration with evolutionary architecture search and research on local learning and plasticity

Availability

Lacuna is under active development. Public release and documentation links will be added here when available.