Construction of Machine Learning Algorithm Based on Fixed Point Theory and Its Application to Wireless Communication Systems

基于不动点理论的机器学习算法构建及其在无线通信系统中的应用

• 批准号: 20760252
• 负责人: YUKAWA Masahiro
• 金额: $ 2.58万
• 依托单位: Niigata University (2010-2011)The Institute of Physical and Chemical Research (2008-2009)
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Young Scientists (B)
• 财政年份: 2008
• 资助国家: 日本
• 起止时间: 2008 至 2011
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-20760252/
• 关键词: 信号処理; 移動体通信; 機械学習; 情報通信工学; 時変計量適応射影劣勾配法; 射影勾配作用素; 複数領域適応学習; 無線通信; 低ランク適応フィルタ

Based on fixed point theory, we have obtained the following outcomes.(1) We have derived a reduced-rank adaptive algorithm, analyzed its convergence properties, and shown that it ameliorates the convergence rate considerably with low computational complexity.(2) We have introduced the new concept of time-varying metric producing a general framework to analyze the convergence properties of a number of adaptive algorithms in a unified fashion.(3) We have imported the concept of feasibility splitting to adaptive signal processing, derived an efficient adaptive algorithm to exploit a variety of information that is expressed in multiple domains, analyzed its convergence properties, and shown its efficacy in SIMO/MIMO wireless communication systems.

基于固定点理论，我们获得了以下结果。（1）我们得出了降低的适应性算法，分析了其收敛属性，并表明它可以大大减轻收敛速率，而计算复杂性低。（2）我们已经介绍了一般框架的新概念，以分析综合属性的新概念。 （3）我们已经将可行性分开的概念导入到自适应信号处理中，得出了一种有效的自适应算法，以利用在多个域中表达的各种信息，分析其收敛性，并在SIMO/MIMO无线通信系统中显示出其功效。

项目成果

How Is Gain of Hybrid Sparse Adaptive Filtering Algorithm Affected by Input Correlation?

输入相关性如何影响混合稀疏自适应滤波算法的增益？

• 发表时间: 2011
• 作者: Osamu Toda;Masahiro Yukawa
• 通讯作者: Masahiro Yukawa

A deterministic analysis of linearly constrained adaptive filtering algorithms

线性约束自适应滤波算法的确定性分析

• 发表时间: 2011
• 作者: Masahiro Yukawa;Isao Yamada
• 通讯作者: Isao Yamada

A Note on Adaptive Projected Subgradient Method Under Linear Constraints

线性约束下自适应投影次梯度法的注记

• 发表时间: 2010
• 作者: M.Yukawa;I.Yamada
• 通讯作者: I.Yamada

On use of multiple kernels in adaptive learning---Extended reproducing kernel Hilbert space with Cartesian product

多核在自适应学习中的应用——带笛卡尔积的扩展再生核希尔伯特空间

• 发表时间: 2010
• 作者: Masao Yamagishi;Masahiro Yukawa;Isao Yamada;Takao Jinno and Masahiro Okuda;Masahiro Yukawa
• 通讯作者: Masahiro Yukawa

On robustness of MIMO antenna selection scheme by minimizing smoothed L1 norm under nonlinear constraints

非线性约束下最小化平滑 L1 范数的 MIMO 天线选择方案的鲁棒性

• 发表时间: 2011
• 作者: Masahiro Yukawa;Isao Yamada
• 通讯作者: Isao Yamada