9:30 AM - 10:00 AM
[K-5-02 (Invited)] Analog Computation-in-Memory (CiM) for Neuromorphic Computing
For edge AI applications, this paper overviews neuromorphic computing with Analog CiM, Computation-in-Memory with non-volatile memories. CiM will be heterogeneously integrated with traditional processors such as CPUs. To extremely suppress energy of edge AI, the heterogeneous integration of sensors like event-based sensors and CiM is promising. In Approximate CiM, by tolerating some degree of device errors, the trade-off of performance, energy and cost is resolved. CiM are effective for various AI algorithms such as Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), event driven Spiking Neural Network (SNN) and Reservoir Computing.