Abstract:The capability for environmental sound recognition (ESR) can determine the fitness of individuals in a way to avoid dangers or pursue opportunities when critical sound events occur. It still remains mysterious about the fundamental principles of biological systems that result in such a remarkable ability. Additionally, the practical importance of ESR has attracted an increasing amount of research attention, but the chaotic and nonstationary difficulties continue to make it a challenging task. In this article, we propose a spike-based framework from a more brain-like perspective for the ESR task. Our framework is a unifying system with consistent integration of three major functional parts which are sparse encoding, efficient learning, and robust readout. We first introduce a simple sparse encoding, where key points are used for feature representation, and demonstrate its generalization to both spike- and nonspike-based systems. Then, we evaluate the learning properties of different learning rules in detail with our contributions being added for improvements. Our results highlight the advantages of multispike learning, providing a selection reference for various spike-based developments. Finally, we combine the multispike readout with the other parts to form a system for ESR. Experimental results show that our framework performs the best as compared to other baseline approaches. In addition, we show that our spike-based framework has several advantageous characteristics including early decision making, small dataset acquiring, and ongoing dynamic processing. Our framework is the first attempt to apply the multispike characteristic of nervous neurons to ESR. The outstanding performance of our approach would potentially contribute to draw more research efforts to push the boundaries of spike-based paradigm to a new horizon.

Online robust echo state broad learning system

Recurrent Broad Learning Systems for Time Series Prediction

TKS-BLS: Temporal Kernel Stationary Broad Learning System for Enhanced Modeling, Anomaly Detection, and Incremental Learning with Application to Ironmaking Processes

A Rank-Based Sampling Framework for Offline Reinforcement Learning

Robust kernel adaptive filtering for nonlinear time series prediction

Robust Environmental Sound Recognition with Sparse Key-point Encoding and Efficient Multi-spike Learning.

A novel broad learning system integrated with restricted Boltzmann machine and echo state network for time series forecasting

Selective Recursive Kernel Learning for Online Identification of Nonlinear Systems with NARX Form

A New Robust Projection Distributed Broad Learning Under Redundant Samples and Noisy Environment

An Ensemble Broad Learning Scheme for Short-Term Load Forecasting

Long-Short Term Echo State Network for Time Series Prediction

Sequential sparse Bayesian learning with applications to system identification for damage assessment and recursive reconstruction of image sequences

Data-driven Robust State Estimation Through Off-line Learning and On-line Matching

Robust Incremental Broad Learning System for Data Streams of Uncertain Scale

Broad learning system based on the quantized minimum error entropy criterion

Robust and Adaptive Online Time Series Prediction with Long Short-Term Memory

Broad Learning With Reinforcement Learning Signal Feedback: Theory and Applications

Spatio-temporal learning with the online finite and infinite echo-state Gaussian processes

Pseudo inverse versus iterated projection: Novel learning approach and its application on broad learning system

Deep Optimized Broad Learning System for Applications in Tabular Data Recognition

Self-organizing deep belief modular echo state network for time series prediction