Papers › Let SSMs be ConvNets: State-space Modeling with Optimal Tensor Contractions

Let SSMs be ConvNets: State-space Modeling with Optimal Tensor Contractions

22 Jan 2025arXiv:2501.13230archive 2025-07-28

Yan Ru Pei

We introduce Centaurus, a class of networks composed of generalized state-space model (SSM) blocks, where the SSM operations can be treated as tensor contractions during training. The optimal order of tensor contractions can then be systematically determined for every SSM block to maximize training efficiency. This allows more flexibility in designing SSM blocks beyond the depthwise-separable configuration commonly implemented. The new design choices will take inspiration from classical convolutional blocks including group convolutions, full convolutions, and bottleneck blocks. We architect the Centaurus network with a mixture of these blocks, to balance between network size and performance, as well as memory and computational efficiency during both training and inference. We show that this heterogeneous network design outperforms its homogeneous counterparts in raw audio processing tasks including keyword spotting, speech denoising, and automatic speech recognition (ASR). For ASR, Centaurus is the first network with competitive performance that can be made fully state-space based, without using any nonlinear recurrence (LSTMs), explicit convolutions (CNNs), or (surrogate) attention mechanism. Source code is available at github.com/Brainchip-Inc/Centaurus

PaperPDF

Code

No code repository is listed for this paper in the archive or in Syntology's graph.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Automatic Speech RecognitionAutomatic Speech Recognition (ASR)Computational EfficiencyDenoisingKeyword SpottingSpeech DenoisingSpeech EnhancementSpeech Recognitionspeech-recognition

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Speech Enhancement VoiceBank + DEMAND Centaurus (0.51M) PESQ (wb) 3.25 #19 of 42 Archive leaderboard report
Speech Recognition LibriSpeech test-clean Centaurus (30 M) Word Error Rate (WER) 4.4 #54 of 64 Archive leaderboard report
Speech Recognition Speech Commands Centaurus Accuracy (%) 98.53 #1 of 3 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

AttentionSoftmax

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections