在尖端神经网络中可学习的轴突延迟可以改善口语识别
Pengfei Sun1, Yansong Chua2, Paul Devos1
1Department of Information Technology, WAVES Research Group, Ghent University, Ghent, Belgium.
Frontiers in neuroscience
|November 29, 2023
概括
尖端神经网络 (SNN) 通过结合可学习的轴突延迟和跳过连接来实现最先进的口语识别. 这种生物启发的方法增强了语言任务的时间处理,以更少的参数优于传统的神经网络.
科学领域:
- 计算神经科学是一种神经科学.
- 人工智能的人工智能
- 语音处理 语音处理
背景情况:
- 尖端神经网络 (SNN) 提供了生物可信的听觉处理模型,对于需要精确时间信息的任务至关重要.
- 与人工神经网络 (ANN) 相比,尖端序列的时间复杂性在历史上限制了SNN的性能.
研究的目的:
- 通过解决尖端定时配置和网络架构方面的挑战,提高SNN在口语识别方面的性能.
- 开发一种新的SNN架构,在具有挑战性的语音基准上取得最先进的结果.
主要方法:
- 在SNN架构中实现了一个可学习的轴突延迟模块.
- 集成的本地跳过连接,以促进信息流和梯度传播.
- 引入了辅助损失术语,以提高模型的准确性和稳定性.
主要成果:
- 在口语识别基准 (NTIDIDIGITS和SHD) 上取得了最先进的表现.
- 显示了显著的性能改进:14%的NTIDIDIGITS和18%的SHD与延迟模块.
- 使用更少的参数 (10x为NTIDIDIGITS,7x为SHD) 的反复和卷积神经网络的表现优于.
结论:
- 可学习的轴突延迟和局部跳过连接在提高语音识别SNN性能方面是有效的.
- 拟议的SNN方法为听觉任务提供了传统深度学习模型的计算效率高和生物灵感替代方案.
相关概念视频
Language Development
369
Children master language quickly and with relative ease, supported by both biological predisposition and reinforcement. B. F. Skinner (1957) proposed that language is learned through reinforcement, while Noam Chomsky (1965) argued that language acquisition mechanisms are biologically determined.
The critical period for language acquisition suggests that the ability to acquire language is at its peak early in life. As people age, this proficiency decreases. Language development begins very...
The critical period for language acquisition suggests that the ability to acquire language is at its peak early in life. As people age, this proficiency decreases. Language development begins very...
369
Long-term Potentiation
55.3K
Long-term potentiation, or LTP, is one of the ways by which synaptic plasticity—changes in the strength of chemical synapses—can occur in the brain. LTP is the process of synaptic strengthening that occurs over time between pre- and postsynaptic neuronal connections. The synaptic strengthening of LTP works in opposition to the synaptic weakening of long-term depression (LTD) and together are the main mechanisms that underlie learning and memory.
55.3K


