COMET-SG1: Lightweight Autoregressive Regressor for Edge and Embedded AI

TL;DR

COMET-SG1 is a lightweight, stable autoregressive model designed for accurate and reliable time-series prediction on edge devices, emphasizing bounded long-term behavior and computational efficiency.

Contribution

It introduces a novel linear, stability-oriented autoregressive architecture optimized for edge AI, outperforming traditional models in long-horizon stability with a compact design.

Findings

Achieves competitive short-horizon accuracy

Exhibits significantly reduced long-horizon drift

Compatible with fixed-point arithmetic for embedded systems

Abstract

COMET-SG1 is a lightweight, stability-oriented autoregressive regression model designed for time-series prediction on edge and embedded AI systems. Unlike recurrent neural networks or transformer-based sequence models, COMET-SG1 operates through linear behavior-space encoding, memory-anchored transition estimation, and deterministic state updates. This structure prioritizes bounded long-horizon behavior under fully autoregressive inference, a critical requirement for edge deployment where prediction errors accumulate over time. Experiments on non-stationary synthetic time-series data demonstrate that COMET-SG1 achieves competitive short-horizon accuracy while exhibiting significantly reduced long-horizon drift compared to MLP, LSTM, and k-nearest neighbor baselines. With a compact parameter footprint and operations compatible with fixed-point arithmetic, COMET-SG1 provides a practical and interpretable approach for stable autoregressive prediction in edge and embedded AI applications.