IMBUE: In-Memory Boolean-to-CUrrent Inference ArchitecturE for Tsetlin Machines

TL;DR

IMBUE is an innovative in-memory computing architecture that leverages ReRAM devices to efficiently perform Boolean-to-current inference for Tsetlin Machines, significantly enhancing performance over traditional digital implementations.

Contribution

The paper introduces IMBUE, a novel ReRAM-based in-memory architecture that eliminates digital-analog conversions for Tsetlin Machines, improving inference speed and efficiency.

Findings

Achieves up to 12.99x speedup over binarized CNNs

Achieves up to 5.28x speedup over digital Tsetlin Machines

Demonstrates effective in-memory Boolean-to-current inference

Abstract

In-memory computing for Machine Learning (ML) applications remedies the von Neumann bottlenecks by organizing computation to exploit parallelism and locality. Non-volatile memory devices such as Resistive RAM (ReRAM) offer integrated switching and storage capabilities showing promising performance for ML applications. However, ReRAM devices have design challenges, such as non-linear digital-analog conversion and circuit overheads. This paper proposes an In-Memory Boolean-to-Current Inference Architecture (IMBUE) that uses ReRAM-transistor cells to eliminate the need for such conversions. IMBUE processes Boolean feature inputs expressed as digital voltages and generates parallel current paths based on resistive memory states. The proportional column current is then translated back to the Boolean domain for further digital processing. The IMBUE architecture is inspired by the Tsetlin Machine (TM), an emerging ML algorithm based on intrinsically Boolean logic. The IMBUE architecture demonstrates significant performance improvements over binarized convolutional neural networks and digital TM in-memory implementations, achieving up to a 12.99x and 5.28x increase, respectively.