Recommending Pre-Trained Models for IoT Devices

TL;DR

This paper introduces a hardware-aware method for recommending pre-trained models tailored for IoT devices, addressing the challenge of selecting suitable models considering resource constraints.

Contribution

It presents a novel, hardware-conscious PTM selection approach and a research agenda for IoT-specific model recommendation systems.

Findings

Identifies limitations of existing model recommendation methods for IoT.

Proposes a new hardware-aware PTM selection technique.

Outlines a research agenda for future IoT model recommendation systems.

Abstract

The availability of pre-trained models (PTMs) has enabled faster deployment of machine learning across applications by reducing the need for extensive training. Techniques like quantization and distillation have further expanded PTM applicability to resource-constrained IoT hardware. Given the many PTM options for any given task, engineers often find it too costly to evaluate each model's suitability. Approaches such as LogME, LEEP, and ModelSpider help streamline model selection by estimating task relevance without exhaustive tuning. However, these methods largely leave hardware constraints as future work-a significant limitation in IoT settings. In this paper, we identify the limitations of current model recommendation approaches regarding hardware constraints and introduce a novel, hardware-aware method for PTM selection. We also propose a research agenda to guide the development of effective, hardware-conscious model recommendation systems for IoT applications.