Search papers, labs, and topics across Lattice.
This paper investigates cumsum-composable phase transport as a novel temporal layer for streaming keyword spotting, addressing the inefficiencies of traditional scan-style training kernels in state-space sequence models. By employing learned unitary rotations and prefix differences, the method achieves competitive accuracy with significantly reduced latency and training time on the Google Speech Commands v2 dataset. Notably, the approach maintains high performance with compact models, achieving up to 97.3% test accuracy while improving efficiency over existing baselines.
Cumsum-composable phase transport enables high-accuracy keyword spotting with a fraction of the model size and training time, revolutionizing streaming speech applications.
State-space sequence models are attractive for streaming speech because they maintain compact recurrent state, but scan-style training kernels can have unfavorable constants for short audio tasks. We study cumsum-composable phase transport, a streaming-native temporal layer for keyword spotting. Each layer projects acoustic frames to complex channels, transports them by learned unitary rotations, accumulates a finite window using prefix differences, and applies a gated residual update. The same prefix representation gives exact batched training with ordinary cumulative sums and exact online inference with one prefix update per frame. Unitary transport is the key constraint: inverse rotations have norm one, keeping prefix terms well conditioned while memory is supplied by windows or block readouts. On Google Speech Commands v2 with 12 labels, mel+cumsum models retain competitive accuracy with compact baselines. The strongest single-seed run reaches 97.3\% test accuracy; a 51.6K-parameter tied model also reaches 97.3\%, and a 24.8K tied model reaches 96.8\% versus 97.1\% for a 25.6K MelCNNMaxPool baseline. In a matched cumsum-versus-scan benchmark, cumsum+window gives comparable accuracy, 94.82\% versus 94.33\%, while training 1.07x faster and reducing single-example latency from 7.09 ms to 5.01 ms on a Tesla T4. These results support cumsum phase transport as a simple low-cost temporal primitive for streaming keyword spotting.
