Adaptive Machine Learning for Resource-Constrained Environments

2025-03-24

Delta

AI Summary

To address the challenges of resource-constrained IoT edge gateways—characterized by dynamic CPU load and continuous data streams—this paper proposes a lightweight online/continual learning framework for time-series forecasting. The framework leverages CPU utilization time-series data and integrates incremental ensemble modeling with adaptive feature representation, enabling real-time performance prediction under zero-shot and fine-tuning settings with low computational overhead and high accuracy. Evaluated on a real-world IoT gateway dataset, it achieves state-of-the-art (SOTA) prediction accuracy while reducing inference latency by over 60% and memory footprint by 5× compared to SOTA baselines such as Lag-Llama. Its core innovation lies in an end-to-end online learning architecture that simultaneously satisfies extreme resource constraints and robust model adaptability.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Planning, Routing, and Scheduling: Planning with Language Models
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

The Internet of Things is an example domain where data is perpetually generated in ever-increasing quantities, reflecting the proliferation of connected devices and the formation of continuous data streams over time. Consequently, the demand for ad-hoc, cost-effective machine learning solutions must adapt to this evolving data influx. This study tackles the task of offloading in small gateways, exacerbated by their dynamic availability over time. An approach leveraging CPU utilization metrics using online and continual machine learning techniques is proposed to predict gateway availability. These methods are compared to popular machine learning algorithms and a recent time-series foundation model, Lag-Llama, for fine-tuned and zero-shot setups. Their performance is benchmarked on a dataset of CPU utilization measurements over time from an IoT gateway and focuses on model metrics such as prediction errors, training and inference times, and memory consumption. Our primary objective is to study new efficient ways to predict CPU performance in IoT environments. Across various scenarios, our findings highlight that ensemble and online methods offer promising results for this task in terms of accuracy while maintaining a low resource footprint.

Problem

• Predicting gateway availability in resource-constrained IoT environments
• Comparing machine learning methods for CPU utilization prediction
• Optimizing accuracy and resource efficiency in dynamic IoT data streams

Innovation

• Online continual learning for gateway prediction
• CPU metrics optimize resource-constrained ML
• Ensemble methods balance accuracy and efficiency