analiza aiqbench
This research introduces Polar MKANs, a novel architecture using Kolmogorov-Arnold Networks to enhance the interpretability of RF fingerprinting systems by explicitly separating magnitude and phase features. By enforcing monotonicity in latent dimensions, the model achieves significantly higher disentanglement scores compared to standard baselines, addressing the opacity issue that hinders deep learning adoption in security-critical wireless authentication.
Interpretable Feature Learning for RF Fingerprinting via Polar MKANs
arXiv:2608.19881v1 Announce Type: cross Abstract: Radio frequency (RF) fingerprinting authenticates wireless devices from hardware-induced I/Q impairments, typically with deep learning feature extractors that are accurate but opaque, limiting their use in security critical settings. We propose Polar Monotonic Kolmogorov-Arnold Networks (Polar MKAN), a block partitioned monotonic encoder on polar inputs in which each latent dimension depends exclusively on magnitude or phase, yielding channel separation and monotone responses by construction. On a synthetic gain and carrier frequency offset (CFO) benchmark, Polar MKAN reaches 57.2 percent DCI Disentanglement versus at most 12.9 percent for unpartitioned baselines. We further evaluate the detection accuracy trade off on real data and the sensitivity to blind CFO compensation.
Abstract:Radio frequency (RF) fingerprinting authenticates wireless devices from hardware-induced I/Q impairments, typically with deep learning feature extractors that are accurate but opaque, limiting their use in security critical settings. We propose Polar Monotonic Kolmogorov-Arnold Networks (Polar MKAN), a block partitioned monotonic encoder on polar inputs in which each latent dimension depends exclusively on magnitude or phase, yielding channel separation and monotone responses by construction. On a synthetic gain and carrier frequency offset (CFO) benchmark, Polar MKAN reaches 57.2 percent DCI Disentanglement versus at most 12.9 percent for unpartitioned baselines. We further evaluate the detection accuracy trade off on real data and the sensitivity to blind CFO compensation.
Signal Processing (eess.SP); Machine Learning (cs.LG)
arXiv:2608.19881 [eess.SP]
(or arXiv:2608.19881v1 [eess.SP] for this version)
https://doi.org/10.48550/arXiv.2608.19881 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 204 cuvinte
Subiecte conexe
Interpretable Physics Informed WiFi Indoor Localization: Learning an Effective Access Point Geometry and Using It to Prune
arXiv · 30 sept. 2026
A systematic investigation of molecular encoding methods for drug property predictions across neural network and Transformer encoder-based model
arXiv · 9 iun. 2026
Multiscale Community-Based Fingerprinting of Signed Functional Networks
arXiv · 31 aug. 2026
Sign-Aware Gated Sparse Autoencoders: Modeling Anticorrelated Features with Bi-Jump-ReLU Activations
arXiv · 4 aug. 2026