The Track on Neuromorphic Software Stacks, Compilers, and Programming Models invites high‑quality submissions that advance the software foundations enabling scalable, efficient, and programmable neuromorphic computing. As neuromorphic hardware platforms diversify—ranging from digital SNN accelerators to analog/mixed‑signal and memristive systems—there is a growing need for robust software abstractions, toolchains, and programming models that bridge the gap between algorithms and hardware. This track brings together researchers from computer architecture, machine learning, systems software, neuromorphic engineering, and programming languages to define the next generation of software infrastructure for brain‑inspired computing.
We welcome original contributions including, but not limited to:
1. Software Stacks and Frameworks
- End‑to‑end software stacks for neuromorphic hardware
- Runtime systems for event‑driven and spiking computation
- APIs and libraries for SNN development, deployment, and profiling
- Integration of neuromorphic systems with mainstream ML frameworks
2. Compilers and Toolchains
- Compiler front‑ends and back‑ends for SNNs and event‑driven workloads
- Mapping, partitioning, and scheduling of large‑scale SNNs
- Code generation for heterogeneous neuromorphic architectures
- Optimization techniques for latency, energy, sparsity, and memory footprint
- Hardware‑aware compilation for analog, mixed‑signal, and emerging‑device systems
3. Programming Models and Abstractions
- High‑level programming models for spiking neural networks
- Domain‑specific languages (DSLs) for neuromorphic computing
- Event‑driven and dataflow‑oriented programming paradigms
- Abstractions for learning rules (STDP, Hebbian, supervised SNN training)
- Models supporting online learning, adaptation, and plasticity
4. Simulation, Debugging, and Verification Tools
- Software simulators for neuromorphic hardware and SNN workloads
- Debugging and visualization tools for spike‑based computation
- Verification and correctness frameworks for neuromorphic systems
- Co‑simulation of hardware, sensors, and real‑time environments
5. System‑Level Integration
- Software–hardware co‑design methodologies
- Interfacing neuromorphic systems with robotics, edge devices, and sensor platforms
- Middleware for distributed and multi‑agent neuromorphic systems
- Real‑time operating systems and event‑driven kernels
6. Applications and Case Studies
- End‑to‑end pipelines for robotics, autonomous systems, and sensorimotor control
- Edge AI applications leveraging neuromorphic software stacks
- Comparative studies of software frameworks across hardware platforms
- Demonstrations of scalable, programmable neuromorphic applications
