Analog AI as a Service: A Cloud Platform for In-Memory Computing

Abstract

This paper introduces the Analog AI Cloud Com- poser platform, a service that allows users to access Ana- log In-Memory Computing (AIMC) simulation and computing resources over the cloud. We also introduce the concept of an Analog AI as a Service (AAaaS). AIMC offers a novel approach for decreasing the latency and energy usage associated with Deep Neural Network (DNN) inference and training. This platform democratizes access to AIMC computing, making it available to a broader audience, including researchers, developers, and businesses. Emphasizing a user-friendly, no-code approach, AAaaS integrates the Analog Hardware Acceleration Kit (AIHWKit) simulation platform within a fully managed cloud environment. The platform caters to researchers, hardware engineers, developers, and AI enthusiasts, enabling them to explore and develop efficient Analog AI neural networks and models through open-source AIHWKit capabilities and an intuitive interface. We discuss the architecture of the Analog AI Cloud Composer (AAICC), focusing on its key services, such as inference, training, and AIMC hardware access. The platform’s design, grounded in cloud services and guidelines, ensures a secure, data-centric user experience with robust control and validation mechanisms. This approach facilitates the creation of AIMC AI models and is an educational tool. The platform is freely accessible at https://aihw-composer. draco.res.ibm.com.