LUT-Compiled Kolmogorov-Arnold Networks for Lightweight DoS Detection on IoT Edge Devices

2026-01-12

AI Summary

This work proposes an efficient real-time DoS attack detection method for resource-constrained IoT edge devices based on Kolmogorov–Arnold Networks (KANs). By deploying learnable univariate B-spline functions on KAN edges, the approach achieves high-accuracy modeling—reaching 99.0% accuracy on CICIDS2017—with an extremely low parameter count of only 50K (0.19 MB). The study introduces lookup table (LUT) compilation into KANs for the first time, combining quantization and linear interpolation to transform spline computations into table lookups. This optimization yields over 5,000× inference speedup at batch size 1 with negligible accuracy degradation (98.96%), substantially enhancing the practicality and deterministic latency performance of KANs in edge computing scenarios.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Denial-of-Service (DoS) attacks pose a critical threat to Internet of Things (IoT) ecosystems, yet deploying effective intrusion detection on resource-constrained edge devices remains challenging. Kolmogorov-Arnold Networks (KANs) offer a compact alternative to Multi-Layer Perceptrons (MLPs) by placing learnable univariate spline functions on edges rather than fixed activations on nodes, achieving competitive accuracy with fewer parameters. However, runtime B-spline evaluation introduces significant computational overhead unsuitable for latency-critical IoT applications. We propose a lookup table (LUT) compilation pipeline that replaces expensive spline computations with precomputed quantized tables and linear interpolation, dramatically reducing inference latency while preserving detection quality. Our lightweight KAN model (50K parameters, 0.19~MB) achieves 99.0\% accuracy on the CICIDS2017 DoS dataset. After LUT compilation with resolution $L=8$, the model maintains 98.96\% accuracy (F1 degradation $<0.0004$) while achieving $\mathbf{68\times}$ speedup at batch size 256 and over $\mathbf{5000\times}$ speedup at batch size 1, with only $2\times$ memory overhead. We provide comprehensive evaluation across LUT resolutions, quantization schemes, and out-of-bounds policies, establishing clear Pareto frontiers for accuracy-latency-memory trade-offs. Our results demonstrate that LUT-compiled KANs enable real-time DoS detection on CPU-only IoT gateways with deterministic inference latency and minimal resource footprint.

Problem

• DoS detection
• IoT edge devices
• lightweight intrusion detection
• resource-constrained systems
• real-time security

Innovation

• Kolmogorov-Arnold Networks
• lookup table compilation
• lightweight DoS detection
• IoT edge security
• spline quantization