Adaptive Random Forests for Energy-Efficient Inference on Microcontrollers

TL;DR

This paper introduces an early-stopping mechanism for Random Forests on microcontrollers, significantly reducing energy consumption for easy inputs while maintaining high accuracy, adaptable at runtime.

Contribution

It presents a novel adaptive RF inference method with runtime-controlled confidence thresholds, improving energy efficiency over existing approaches.

Findings

Energy reduction of 38% to over 90% on embedded tasks

Less than 0.5% accuracy loss with energy savings

Outperforms previous adaptive RF methods

Abstract

Random Forests (RFs) are widely used Machine Learning models in low-power embedded devices, due to their hardware friendly operation and high accuracy on practically relevant tasks. The accuracy of a RF often increases with the number of internal weak learners (decision trees), but at the cost of a proportional increase in inference latency and energy consumption. Such costs can be mitigated considering that, in most applications, inputs are not all equally difficult to classify. Therefore, a large RF is often necessary only for (few) hard inputs, and wasteful for easier ones. In this work, we propose an early-stopping mechanism for RFs, which terminates the inference as soon as a high-enough classification confidence is reached, reducing the number of weak learners executed for easy inputs. The early-stopping confidence threshold can be controlled at runtime, in order to favor either energy saving or accuracy. We apply our method to three different embedded classification tasks, on a single-core RISC-V microcontroller, achieving an energy reduction from 38% to more than 90% with a drop of less than 0.5% in accuracy. We also show that our approach outperforms previous adaptive ML methods for RFs.