We are proud to announce that the paper titled "When Audio Denoising Meets Spiking Neural Network", authored by our PhD students, Mr HAO Xiang and Mr MA Chenxiang, and collaborated with Miss Qu YANG, National University of Singapore; Prof. Tan Kay Chen, Interim Head and Chair Professor of Computational Intelligence, DSAI, PolyU; and Dr Wu Jibin, Assistant Professor, COMP, received the Best Paper Award at the 2024 IEEE Conference on Artificial Intelligence (IEEE CAI 2024) held in Singapore from June 25 - 27, 2024.
The award-winning paper explores the application of spiking neural networks (SNNs) in audio denoising, leveraging brain-inspired architectures and computational principles to efficiently process and analyse audio signals. The paper introduces Spiking-FullSubNet, a cutting-edge, real-time audio denoising model based on SNN. The model incorporates a novel gated spiking neuron model (GSN) to effectively capture multi-scale temporal information, which is essential for achieving high-fidelity audio denoising. In addition to the architectural advancements, the authors propose the integration of GSNs within an optimised FullSubNet neural architecture, enabling efficient processing of full-band and sub-band frequencies while significantly reducing computational overhead. The paper also introduces a metric discriminator-based loss function that selectively enhances the desired performance metrics without compromising others. Empirical evaluations demonstrate the superior performance of Spiking-FullSubNet, which ranked as the winner of Track 1 (Algorithmic) of the Intel Neuromorphic Deep Noise Suppression Challenge.
The IEEE CAI 2024 is a prominent industry-organized event dedicated to artificial intelligence and its applications in various industries. The conference serves as a platform for connecting AI enterprise leaders, innovators, and stakeholders from industry, government, and start-ups, showcasing the latest advancements in AI research and applications.
