Compute-in-memory implementation of state space models for event sequence processing

State space models have recently emerged as a powerful framework for long sequence processing, outperforming traditional methods on diverse benchmarks. Fundamentally, state space models can generalize both recurrent and convolutional networks and have been shown to even capture key functions of biological systems. Here we report an approach to implement SSMs in energy-efficient compute-in-memory hardware to achieve real-time, event-driven processing. Our work re-parameterizes the model to function with real-valued coefficients and shared decay constants, reducing the complexity of model mapping onto practical hardware systems. By leveraging device dynamics and diagonalized state transition parameters, the state evolution can be natively implemented in crossbar-based compute-in-memory systems combined with memristors exhibiting short-term memory effects. Through this algorithm and hardware co-design, we show the proposed system offers both high accuracy and high energy efficiency while supporting fully asynchronous processing for event-based vision and audio tasks. Spiking neural networks are generally used for sequential information and event data processing but still lack high performance. Through algorithm and hardware co-design, Zhang et al. report a State Space Model based approach to implement on compute-in memory hardware, enabling asynchronous and real-time processing capability with high energy efficiency for event sequences.