We welcome focused collaborations built around an explicit question, common baselines and a result that another group can reproduce.
Approximation and universality in generalized Walsh systems; operator boundedness and compactness; discretization and recovery with explicit function-space assumptions.
A precise statement, counterexample analysis and independently checked proof.
Common-budget Haar–Walsh, DWT/SWT, DCT/FFT and learned baselines. Separate feature information from classifier and tuning effects.
A shared protocol, reproducible implementations and uncertainty-aware comparisons.
ECG acquisition failures, missing samples and corruption. Audit patient identities, annotations and the information available at prediction time.
A frozen benchmark with error, coverage and failure-mode reports.
Complete processing measurements on ARM, from detection and preprocessing to inference. Compare model storage, memory, buffering and measured latency.
A runnable prototype, versioned configurations and device measurement logs.
