Out-of-Distribution Detection via Channelwise Feature Aggregation in Neural Network-Based Receivers
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911375619522560 |
|---|---|
| author | Tuononen, Marko Penttinen, Heikki Vu, Duy Korpi, Dani Starck, Vesa Hautamäki, Ville |
| author_facet | Tuononen, Marko Penttinen, Heikki Vu, Duy Korpi, Dani Starck, Vesa Hautamäki, Ville |
| contents | Neural network-based radio receivers are expected to play a key role in future wireless systems, making reliable Out-Of-Distribution (OOD) detection essential. We propose a post-hoc, layerwise OOD framework based on channelwise feature aggregation that avoids classwise statistics--critical for multi-label soft-bit outputs with astronomically many classes. Receiver activations exhibit no discrete clusters but a smooth Signal-to-Noise-Ratio (SNR)-aligned manifold, consistent with classical receiver behavior and motivating manifold-aware OOD detection. We evaluate multiple OOD feature types, distance metrics, and methods across layers. Gaussian Mahalanobis with mean activations is the strongest single detector, earlier layers outperform later, and SNR/classifier fusions offer small, inconsistent AUROC gains. High-delay OOD is detected reliably, while high-speed remains challenging. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_15570 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Out-of-Distribution Detection via Channelwise Feature Aggregation in Neural Network-Based Receivers Tuononen, Marko Penttinen, Heikki Vu, Duy Korpi, Dani Starck, Vesa Hautamäki, Ville Machine Learning Signal Processing 68T07 (Primary) 62H30, 62G05, 94A05, 68T30 (Secondary) I.2.6; I.5.1; C.2.3 Neural network-based radio receivers are expected to play a key role in future wireless systems, making reliable Out-Of-Distribution (OOD) detection essential. We propose a post-hoc, layerwise OOD framework based on channelwise feature aggregation that avoids classwise statistics--critical for multi-label soft-bit outputs with astronomically many classes. Receiver activations exhibit no discrete clusters but a smooth Signal-to-Noise-Ratio (SNR)-aligned manifold, consistent with classical receiver behavior and motivating manifold-aware OOD detection. We evaluate multiple OOD feature types, distance metrics, and methods across layers. Gaussian Mahalanobis with mean activations is the strongest single detector, earlier layers outperform later, and SNR/classifier fusions offer small, inconsistent AUROC gains. High-delay OOD is detected reliably, while high-speed remains challenging. |
| title | Out-of-Distribution Detection via Channelwise Feature Aggregation in Neural Network-Based Receivers |
| topic | Machine Learning Signal Processing 68T07 (Primary) 62H30, 62G05, 94A05, 68T30 (Secondary) I.2.6; I.5.1; C.2.3 |
| url | https://arxiv.org/abs/2505.15570 |