23 February 2022 in Computational Intelligence and Neuroscience
The swarm intelligence algorithm is a new technology proposed by researchers inspired by the biological behavior of nature, which has been practically applied in various fields. As a kind of swarm intelligence algorithm, the newly proposed sparrow search algorithm has attracted extensive attention due to its strong optimization ability. Aiming at the problem that it is easy to fall into local optimum, this paper proposes an improved sparrow search algorithm (IHSSA) that combines infinitely folded iterative chaotic mapping (ICMIC) and hybrid reverse learning strategy. In the population initialization stage, the improved ICMIC strategy is combined to increase the distribution breadth of the population and improve the quality of the initial solution. In the finder update stage, a reverse learning strategy based on the lens imaging principle is utilized to update the group of discoverers with high fitness, while the generalized reverse learning strategy is used to update the current global worst solution in the joiner update stage. To balance exploration and exploitation capabilities, crossover strategy is joined to update scout positions. 14 common test functions are selected for experiments, and the Wilcoxon rank sum test method is achieved to verify the effect of the algorithm, which proves that IHSSA has higher accuracy and better convergence performance to obtain solutions than 9 algorithms such as WOA, GWO, PSO, TLBO, and SSA variants. Finally, the IHSSA algorithm is applied to three constrained engineering optimization problems, and satisfactory results are held, which proves the effectiveness and feasibility of the improved algorithm.
8 January 2026 in Neural Networks
1 March 2025 in Neuromorphic Computing and Engineering
Finite state machines (FSMs) are fundamental models widely used in a variety of domains for state control. However, they face challenges in modeling complex especially black-box systems without explicit state descriptions. Neural networks, conversely, excel at modeling implicit and continuous systems but struggling with the temporally stable and precise tasks which FSMs can handle effectively. This work explores the emulation of FSMs with neural networks to harness the strengths of both paradigms. Inspired by the similarity between the spike-based information representation in bio-inspired spiking neural networks (SNNs) and the discrete state transition in FSMs, we propose discrete-time spiking recurrent neural networks (DTSRNNs) to emulate FSMs. We further incorporate one-hot encoding to enhance the discriminability of state vectors, which is beneficial for learning complex behaviors. Then, we build a random-FSM dataset to evaluate model performance. Extensive experiments reveal that, our DTSRNNs surpass conventional discrete-time recurrent neural networks (DTRNNs) with extended decline periods, indicating superior temporal stability. They also exhibit higher robustness against different types of noise. Our work not only presents a significant advancement in the stable emulation of general FSMs with SNNs, but also provides a promising approach for modeling complex temporal systems with especially long sequences.