Semantic Meta-Split Learning: A TinyML Scheme for Few-Shot Wireless Image Classification

TL;DR

This paper introduces a TinyML-based semantic communication framework combining split-learning and meta-learning for efficient, privacy-preserving few-shot wireless image classification, achieving significant accuracy gains with less data and energy.

Contribution

It proposes a novel semantic meta-split learning scheme that integrates split-learning and meta-learning for TinyML, addressing computational, data, and privacy challenges.

Findings

Achieves 20% higher classification accuracy over conventional methods.

Requires fewer data points and less training energy.

Provides uncertainty analysis using conformal prediction techniques.

Abstract

Semantic and goal-oriented (SGO) communication is an emerging technology that only transmits significant information for a given task. Semantic communication encounters many challenges, such as computational complexity at end users, availability of data, and privacy-preserving. This work presents a TinyML-based semantic communication framework for few-shot wireless image classification that integrates split-learning and meta-learning. We exploit split-learning to limit the computations performed by the end-users while ensuring privacy-preserving. In addition, meta-learning overcomes data availability concerns and speeds up training by utilizing similarly trained tasks. The proposed algorithm is tested using a data set of images of hand-written letters. In addition, we present an uncertainty analysis of the predictions using conformal prediction (CP) techniques. Simulation results show that the proposed Semantic-MSL outperforms conventional schemes by achieving 20 % gain on classification accuracy using fewer data points, yet less training energy consumption.