Environmental sounds form part of our daily life. With the advancement of deep learning models and the abundance of training data, the performance of automatic sound classification (ASC) systems has improved significantly in recent years. However, the high computational cost, hence high power consumption, remains a major hurdle for large-scale implementation of ASC systems on mobile and wearable devices. Motivated by the observations that humans are highly effective and consume little power whilst analyzing complex audio scenes, we propose a biologically plausible ASC framework, namely SOM-SNN. This framework uses the unsupervised self-organizing map (SOM) for representing frequency contents embedded within the acoustic signals, followed by an event-based spiking neural network (SNN) for spatiotemporal spiking pattern classification. We report experimental results on the RWCP environmental sound and TIDIGITS spoken digits datasets, which demonstrate competitive classification accuracies over other deep learning and SNN-based models. The SOM-SNN framework is also shown to be highly robust to corrupting noise after multi-condition training, whereby the model is trained with noise-corrupted sound samples. Moreover, we discover the early decision making capability of the proposed framework: an accurate classification can be made with an only partial presentation of the input.

Spiking Neural Networks (SNNs) use spatiotemporal spike patterns to represent and transmit information, which are not only biologically realistic but also suitable for ultralow-power event-driven neuromorphic implementation. Just like other deep learning techniques, Deep Spiking Neural Networks (DeepSNNs) benefit from the deep architecture. However, the training of DeepSNNs is not straightforward because the wellstudied error back-propagation (BP) algorithm is not directly applicable. In this paper, we first establish an understanding as to why error back-propagation does not work well in DeepSNNs. We then propose a simple yet efficient Rectified Linear Postsynaptic Potential function (ReL-PSP) for spiking neurons and a Spike-Timing-Dependent Back-Propagation (STDBP) learning algorithm for DeepSNNs where the timing of individual spikes is used to convey information (temporal coding), and learning (back-propagation) is performed based on spike timing in an event-driven manner. We show that DeepSNNs trained with the proposed single spike time-based learning algorithm can achieve state-of-the-art classification accuracy. Furthermore, by utilizing the trained model parameters obtained from the proposed STDBP learning algorithm, we demonstrate ultra-lowpower inference operations on a recently proposed neuromorphic inference accelerator. The experimental results also show that the neuromorphic hardware consumes 0.751 mW of the total power consumption and achieves a low latency of 47.71 ms to classify an image from the MNIST dataset. Overall, this work investigates the contribution of spike timing dynamics for information encoding, synaptic plasticity and decision making, providing a new perspective to the design of future DeepSNNs and neuromorphic hardware.

No abstract

BackgroundTo evaluate the risk of the recurrence and the efficiency of the vaccination, we followed-up antibody responses in patients with the 2009 pandemic H1N1 influenza and persons who received the pandemic H1N1 vaccine in Guangzhou China.MethodsWe collected serum samples from 129 patients and 86 vaccinated persons at day 0, 15, 30, 180 after the disease onset or the vaccination, respectively. Antibody titers in these serum samples were determined by haemagglutination inhibition (HI) assay using a local isolated virus strain A/Guangdong Liwan/SWL1538/2009(H1N1).ResultsHI antibody positive rate of the patients increased significantly from 0% to 60% at day 15 (χ2 = 78, P<0.001) and 100% at day 30 (χ2 = 23, P<0.001), but decreased significantly to 52% at day 180 (χ2 = 38, P<0.001), while that of vaccinated subjects increased from 0% to 78% at day 15 (χ2 = 110, P<0.001) and 81% at day 30 (χ2 = 0.32, P = 0.57), but decreased significantly to 34% at day 180 (χ2 = 39, P<0.001). Geometric mean titers (GMT) of HI antibodies in positive samples from the patients did not change significantly between day 15 and day 30 (T = 0.92, P = 0.36), but it decreased significantly from 80 at day 30 to 52 at day 180 (T = 4.5, P<0.001). GMT of vaccinated persons increased significantly from 100 at day 15 to 193 at day 30 (T = 4.5, P<0.001), but deceased significantly to 74 at day 180 (T = 5.1, P<0.001). Compared to the patients, the vaccinated subjects showed lower seroconversion rate (χ2 = 11, P<0.001; χ2 = 5.9, P = 0.015), but higher GMT (T = 6.0, P<0.001; T = 3.6, P = 0.001) at day 30 and day 180, respectively.ConclusionVaccination of 2009 influenza A (H1N1) was effective. However, about half or more recovered patients and vaccinated persons might have lost sufficient immunity against the recurrence of the viral infection after half a year. Vaccination or re-vaccination may be necessary for prevention of the recurrence.

No abstract