一个集成Mamba和Cgmlp的Mcformer编码器,用于改进声学特征提取
Nurmemet Yolwas1,2, Yongchao Li3,4, Lixu Sun3,4
1School of Computer Science and Technology (School of Cyberspace Security), Xinjiang University, Urumqi, 830046, China. nurmemet@xju.edu.cn.
Scientific reports
|July 2, 2025
概括
新型的Mcformer编码器通过使用Mamba模块增强全球上下文处理和使用Convolutional Gated Multilayer Perceptrons进行本地特征提取来提高语音识别. 这种新模型显著降低了多个公共数据集的字符和文字错误率.
科学领域:
- 人工智能的人工智能
- 计算机科学 计算机科学
- 信号处理 信号处理
背景情况:
- 符合架构是语音识别的主流,集成自我注意力和卷积网络.
- 符合者在捕获全球信息方面存在局限性,阻碍了语音识别性能.
- 现有的模型需要进一步优化,以进行全面的特征提取.
研究的目的:
- 引入Mcformer编码器以提高语音识别能力.
- 在语音识别模型中增强全球上下文处理和本地特征提取.
- 解决当前注意力模型在捕获全面的上下文信息方面的局限性.
主要方法:
- 麦克福默的编码器结合了Mamba模块并行与多头注意力块.
- 为了改进局部特征提取,采用了卷积门式多层感知器 (Cgmlp) 结构.
- 该模型在Aishell-1,Common Voice zh 14和TED-LIUM 3数据集上进行了评估.
主要成果:
- 在Aishell-1验证和测试集中,Mcformer的字符错误率 (CER) 分别为4.15%和4.48% (没有语言模型).
- 在Common Voice zh 14验证和测试集上,Mcformer分别实现了13.28%和13.06%的CER (没有语言模型).
- 在TED-LIUM 3数据集上,Mcformer在验证和测试集 (没有语言模型) 上分别实现了7.26%和6.95%的文字错误率 (WER).
结论:
- 麦克福默尔编码器在语音识别准确度方面取得了显著的改进.
- 整合Mamba模块和Cgmlp结构有效地增强了全球和本地特征处理.
- 实验结果验证了Mcformer在减少不同数据集和语言的错误率方面的有效性.
相关概念视频
00:56Extraction: Advanced Methods
547
Metal ions can be separated from one another by complexation with organic ligands–the chelating agent– to form uncharged chelates. Here, the chelating agent must contain hydrophobic groups and behave as a weak acid, losing a proton to bind with the metal. Since most organic ligands used in this process are insoluble or undergo oxidation in the aqueous phase, the chelating agent is initially added to the organic phase and extracted into the aqueous phase. The metal-ligand complex is...
547
00:58¹H NMR Signal Integration: Overview
1.7K
The intensity of a signal, which can be represented by the area under the peak, depends on the number of protons contributing to that signal. The area under each peak is shown as a vertical line called an integral, with the integral value listed under it, as seen in the proton NMR spectrum of benzyl acetate. Each integral value is divided by the smallest integral value to obtain the ratio of the number of protons producing each signal. The ratio reveals the relative number of protons and not...
1.7K
01:14Extraction: Partition and Distribution Coefficients
3.0K
The distribution law or Nernst's distribution law is the law that governs the distribution of a solute between two immiscible solvents. This law, also known as the partition law, states that if a solute is added to the mixture of two immiscible solvents at a constant temperature, the solute is distributed between the two solvents in such a way that the ratio of solute concentrations in the solvents remains constant at equilibrium.
For extracting a solute from an aqueous phase into an...
For extracting a solute from an aqueous phase into an...
3.0K
01:10Reconstruction of Signal using Interpolation
353
Signal processing techniques are essential for accurately converting continuous signals to digital formats and vice versa. When a continuous signal is sampled with a period T, the resulting sampled signal exhibits replicas of the original spectrum in the frequency domain, spaced at intervals equal to the sampling frequency. To handle this sampled signal, a zero-order hold method can be applied, which creates a piecewise constant signal by retaining each sample's value until the next...
353
01:26Linear Approximation in Frequency Domain
137
Linear systems are characterized by two main properties: superposition and homogeneity. Superposition allows the response to multiple inputs to be the sum of the responses to each individual input. Homogeneity ensures that scaling an input by a scalar results in the response being scaled by the same scalar.
In contrast, nonlinear systems do not inherently possess these properties. However, for small deviations around an operating point, a nonlinear system can often be approximated as linear....
In contrast, nonlinear systems do not inherently possess these properties. However, for small deviations around an operating point, a nonlinear system can often be approximated as linear....
137
01:20Deconvolution
262
Deconvolution, also known as inverse filtering, is the process of extracting the impulse response from known input and output signals. This technique is vital in scenarios where the system's characteristics are unknown, and they must be inferred from the observable signals.
Deconvolution involves several mathematical techniques to derive the impulse response. One common approach is polynomial division. In this method, the input and output sequences are treated as coefficients of...
Deconvolution involves several mathematical techniques to derive the impulse response. One common approach is polynomial division. In this method, the input and output sequences are treated as coefficients of...
262