Deep Convolutional Spiking Neural Networks for Keyword Spotting
This paper investigates the use of deep convolutional spiking neural networks (SNN) for keyword spotting (KWS) and wakeword detection tasks. The brain-inspired SNN mimic the spike-based information processing of biological neural networks and they can operate on the emerging ultra-low power neuromorphic chips. Unlike conventional artificial neural networks (ANN), SNN process input information asynchronously in an event-driven manner. With temporally sparse input information, this event-driven processing substantially reduces the computational requirements compared to the synchronous computation performed in ANN-based KWS approaches. To explore the effectiveness and computational complexity of SNN on KWS and wakeword detection, we compare the performance and computational costs of spiking fully-connected and convolutional neural networks with ANN counterparts under clean and noisy testing conditions. The results obtained on the Speech Commands and Hey Snips corpora have shown the effectiveness of the convolutional SNN model compared to a conventional CNN with comparable performance on KWS and better performance on the wakeword detection task. With its competitive performance and reduced computational complexity, convolutional SNN models running on energy-efficient neuromorphic hardware offer a low-power and effective solution for mobile KWS applications.
Code (0)
등록된 구현이 없습니다.
Tasks
Keyword SpottingSimilar Papers 제목 키워드 기반
WaveSense: Efficient Temporal Convolutions with Spiking Neural Networks for Keyword Spotting
Ultra-low power local signal processing is a crucial aspect for edge applications on always-on devices. Neuromorphic processors emulating spiking neural networks show great computational power while fulfilling the limite…
Keyword SpottingA Comparison of Temporal Encoders for Neuromorphic Keyword Spotting with Few Neurons
With the expansion of AI-powered virtual assistants, there is a need for low-power keyword spotting systems providing a "wake-up" mechanism for subsequent computationally expensive speech recognition. One promising appro…
Binary ClassificationKeyword Spottingspeech-recognitionSpeech RecognitionNeuromorphic Keyword Spotting with Pulse Density Modulation MEMS Microphones
The Keyword Spotting (KWS) task involves continuous audio stream monitoring to detect predefined words, requiring low energy devices for continuous processing. Neuromorphic devices effectively address this energy challen…
Keyword SpottingEncoder-Decoder Neural Architecture Optimization for Keyword Spotting
Keyword spotting aims to identify specific keyword audio utterances. In recent years, deep convolutional neural networks have been widely utilized in keyword spotting systems. However, their model architectures are mainl…
DecoderKeyword SpottingNeural Architecture SearchGlobal-Local Convolution with Spiking Neural Networks for Energy-efficient Keyword Spotting
Thanks to Deep Neural Networks (DNNs), the accuracy of Keyword Spotting (KWS) has made substantial progress. However, as KWS systems are usually implemented on edge devices, energy efficiency becomes a critical requireme…
Keyword Spotting