NeuroAI Temporal Neural Networks (NeuTNNs): Microarchitecture and Design Framework for Specialized Neuromorphic Processing Units

2026-02-02

AI Summary

This work addresses next-generation neuromorphic computing by bridging neuroscience mechanisms with artificial intelligence, balancing biological plausibility and hardware efficiency. It proposes NeuroAI Temporal Neural Networks (NeuTNNs), a novel architecture that, for the first time, integrates neuron models featuring active dendrites and proximal-distal compartmentalization into spiking neural networks, along with a customizable NeuTNN microarchitecture. To enable efficient algorithm-to-hardware mapping, the authors develop NeuTNNGen, a PyTorch-to-layout toolchain incorporating synaptic pruning to reduce resource overhead. Experimental results demonstrate that NeuTNNs outperform existing temporal neural network approaches in both accuracy and energy efficiency on UCR time-series, MNIST, and Place Cells benchmarks. Furthermore, the pruning strategy achieves 30–50% hardware cost reduction without compromising performance.

Technology Category

Cognitive Modeling & Cognitive Systems: Neural Spike CodingMachine Learning: Hardware-aware MLHumans and AI: Other Foundations of Human Computation & AI

Application Category

Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphsSystems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applicationsEconomics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

Leading experts from both communities have suggested the need to (re)connect research in neuroscience and artificial intelligence (AI) to accelerate the development of next-generation AI innovations. They term this convergence as NeuroAI. Previous research has established temporal neural networks (TNNs) as a promising neuromorphic approach toward biological intelligence and efficiency. We fully embrace NeuroAI and propose a new category of TNNs we call NeuroAI TNNs (NeuTNNs) with greater capability and hardware efficiency by adopting neuroscience findings, including a neuron model with active dendrites and a hierarchy of distal and proximal segments. This work introduces a PyTorch-to-layout tool suite (NeuTNNGen) to design application-specific NeuTNNs. Compared to previous TNN designs, NeuTNNs achieve superior performance and efficiency. We demonstrate NeuTNNGen's capabilities using three example applications: 1) UCR time series benchmarks, 2) MNIST design exploration, and 3) Place Cells design for neocortical reference frames. We also explore using synaptic pruning to further reduce synapse counts and hardware costs by 30-50% while maintaining model precision across diverse sensory modalities. NeuTNNGen can facilitate the design of application-specific energy-efficient NeuTNNs for the next generation of NeuroAI computing systems.

Problem

• NeuroAI
• Temporal Neural Networks
• Neuromorphic Computing
• Hardware Efficiency
• Application-specific Design

Innovation

• NeuroAI
• Temporal Neural Networks
• Active Dendrites
• Neuromorphic Computing
• Synaptic Pruning