SensorBench: Benchmarking LLMs in Coding-Based Sensor Processing

TL;DR

SensorBench evaluates the capabilities of Large Language Models in processing sensor data, revealing their strengths and limitations across diverse real-world tasks, and explores prompting strategies to enhance performance.

Contribution

This paper introduces SensorBench, a comprehensive benchmark for assessing LLMs in sensor data processing, and analyzes prompting strategies to improve their effectiveness.

Findings

LLMs perform well on simple sensor tasks

Challenges remain in complex, compositional tasks

Self-verification prompts outperform other methods in 48% of cases

Abstract

Effective processing, interpretation, and management of sensor data have emerged as a critical component of cyber-physical systems. Traditionally, processing sensor data requires profound theoretical knowledge and proficiency in signal-processing tools. However, recent works show that Large Language Models (LLMs) have promising capabilities in processing sensory data, suggesting their potential as copilots for developing sensing systems. To explore this potential, we construct a comprehensive benchmark, SensorBench, to establish a quantifiable objective. The benchmark incorporates diverse real-world sensor datasets for various tasks. The results show that while LLMs exhibit considerable proficiency in simpler tasks, they face inherent challenges in processing compositional tasks with parameter selections compared to engineering experts. Additionally, we investigate four prompting strategies for sensor processing and show that self-verification can outperform all other baselines in 48% of tasks. Our study provides a comprehensive benchmark and prompting analysis for future developments, paving the way toward an LLM-based sensor processing copilot.