Synthetic User Behavior Sequence Generation with Large Language Models for Smart Homes

TL;DR

The paper tackles the challenge that smart-home security models trained on fixed datasets struggle to generalize to evolving user behaviors and environments while raising privacy concerns. It introduces IoTGen, a two-module framework combining Structure Pattern Perception Compression (SPPC) for token-efficient sequence representation and a large language model–driven IoT synthetic data generator that uses structured prompts to produce scene-aware data. SPPC ranks and preserves informative behavior sequences based on reconstruction loss, enabling compression with minimal loss of predictive information, while the generator converts compressed sequences into text and expands them into new scenes via role, task, scene, and data instructions. The proposed approach aims to improve generalization of downstream anomaly detection and behavior-prediction tasks, reduce real-data collection, and enable privacy-preserving, open-world smart-home systems with adaptable security models.

Abstract

In recent years, as smart home systems have become more widespread, security concerns within these environments have become a growing threat. Currently, most smart home security solutions, such as anomaly detection and behavior prediction models, are trained using fixed datasets that are precollected. However, the process of dataset collection is time-consuming and lacks the flexibility needed to adapt to the constantly evolving smart home environment. Additionally, the collection of personal data raises significant privacy concerns for users. Lately, large language models (LLMs) have emerged as a powerful tool for a wide range of tasks across diverse application domains, thanks to their strong capabilities in natural language processing, reasoning, and problem-solving. In this paper, we propose an LLM-based synthetic dataset generation IoTGen framework to enhance the generalization of downstream smart home intelligent models. By generating new synthetic datasets that reflect changes in the environment, smart home intelligent models can be retrained to overcome the limitations of fixed and outdated data, allowing them to better align with the dynamic nature of real-world home environments. Specifically, we first propose a Structure Pattern Perception Compression (SPPC) method tailored for IoT behavior data, which preserves the most informative content in the data while significantly reducing token consumption. Then, we propose a systematic approach to create prompts and implement data generation to automatically generate IoT synthetic data with normative and reasonable properties, assisting task models in adaptive training to improve generalization and real-world performance.

Figures (4)

• Figure 1: The Overview of IoTGen.
• Figure 2: Reconstruction Loss of Model Trained with Full/Compressed Data on The Test Dataset (TOP 50).
• Figure 3: Mean of Loss with Model Trained at Different Compression Levels on The Test Dataset.
• Figure 4: Variance of Loss with Model Trained at Different Compression Levels on The Test Dataset.