Graph neural networks for sampling-invariant embeddings of organized signal sets

The paper proposes graph neural network encoders that transform organized signal sets—such as ordered radar or sensor signals with varying sampling parameters—into fixed‑size vector representations. These embeddings aim to eliminate sampling differences, enabling topology‑aware processing without traditional signal‑processing constraints. The authors evaluate the discriminative power of the embeddings, focusing on waveform separation, using synthetic complex‑valued radiofrequency signals.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

• rag