大规模时间序列表示学习通过同时低频和高频特征启动
IEEE transactions on neural networks and learning systems
|November 28, 2023
概括
本研究介绍了一种新的非对比自主监督学习 (SSL) 方法,用于时间序列表示学习. 该方法有效地捕捉了低频和高频特征,提高了概括性和降低了计算成本.
科学领域:
- 机器学习 机器学习
- 时间序列分析时间序列分析
- 深度学习 (Deep Learning) 是一种深度学习.
背景情况:
- 从未标记的时间序列数据中学习表示具有挑战性.
- 现有的方法难以同时捕捉低频和高频特征,限制了概括性.
- 许多当前的方法使用大型模型或计算上昂贵的技术,如对比学习.
研究的目的:
- 为时间序列表示学习提出一种高效,非对比的自我监督学习 (SSL) 方法.
- 以成本有效的方式有效捕捉低频和高频特征.
- 提高学习表征的概括能力.
主要方法:
- 一个语神经网络配置,有两个重量共享分支.
- 使用多层感知器 (MLP) 和时间卷积网络 (TCN) 头的低频和高频特征提取模块.
- 从单个集合中使用随机转换来增加输入数据.
主要成果:
- 拟议的方法有效地捕获低频和高频特征.
- 在五个真实世界时间序列数据集上实现了最先进的性能.
- 通过广泛的实验和废弃研究证明了强度.
结论:
- 开发的非对比的SSL框架有效地学习了强大的时间序列表示.
- 该方法为捕获多频特征提供了具有成本效益的解决方案.
- 在时间序列表示学习中优于现有的方法.
相关概念视频
Bootstrapping
610
The term "bootstrap" originated in the 19th century as a metaphor for self-improvement or achieving something independently, without external assistance. This concept extends to statistical bootstrapping, a self-contained method for estimating population parameters through resampling, even though it can be computationally intensive. Developed by the American statistician Dr. Bradley Efron in 1979, bootstrapping provides a robust way to perform inference when the original sample size is...
610
Linear Approximation in Frequency Domain
92
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....
92
Aliasing
140
Accurate signal sampling and reconstruction are crucial in various signal-processing applications. A time-domain signal's spectrum can be revealed using its Fourier transform. When this signal is sampled at a specific frequency, it results in multiple scaled replicas of the original spectrum in the frequency domain. The spacing of these replicas is determined by the sampling frequency.
If the sampling frequency is below the Nyquist rate, these replicas overlap, preventing the original...
If the sampling frequency is below the Nyquist rate, these replicas overlap, preventing the original...
140
Linear Approximation in Time Domain
83
Nonlinear systems often require sophisticated approaches for accurate modeling and analysis, with state-space representation being particularly effective. This method is especially useful for systems where variables and parameters vary with time or operating conditions, such as in a simple pendulum or a translational mechanical system with nonlinear springs.
For a simple pendulum with a mass evenly distributed along its length and the center of mass located at half the pendulum's length,...
For a simple pendulum with a mass evenly distributed along its length and the center of mass located at half the pendulum's length,...
83
Time-Series Graph
4.4K
A time-series graph is a line graph with repeated measurements taken at successive intervals of time. It is also called a time series chart. To construct a time-series graph, one must look at both pieces of a paired data set. The horizontal axis is used to plot the time increments, and the vertical axis is used to plot the values of the variable that one is measuring. By using the axes in this way, each point on the graph will correspond to time and a measured quantity. The points on the graph...
4.4K
Reconstruction of Signal using Interpolation
207
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...
207


