… with an infinite number of possible choices of words and stories, and with no clear endpoint.
As such, it should benefit from any change that enhances nonlinear, creative functions. Thus…

: 磁异常化极是磁测数据处理的重要基础工作, 低纬度磁异常化极运算不稳定. 深度学习是一种
有效解决不稳定问题的数据驱动方法. 然而, 现有基于深度学习的位场数据处理仅以位场响应作为…

: 癫痫是一种严重的且反复发作的多因素的神经系统疾病. 对于癫痫的准确诊断和治疗需要对其
发病机制有全面的了解. 研究表明, 癫痫的发病机制复杂, 涉及基因表达, 蛋白质表达, 离子通道…

: 理解并预测多尺度, 高维度和非线性的地震学现象是一个极具挑战性的科学任务. 与日俱增的
海量观测数据降低了信息收集和信息解读之间的耦合程度, 增加了信息解读的抽象性和不确定性. …

本篇論文的研究目的是希望能製做出一系統,能使童伴機器人與孩童進行互動,並從中學習單字.系統
被設計成三個步驟,包含抽取物件的區域,抽取字元以及字元辨識.本篇論文有主要兩個貢獻,第…

… In this thesis we report two new application cases for the Artificial Neuron Network they are
focus on the laser plus characterization and the locating of light spot. By using the learning …

… On pixel-wise explanations for non-linear classifier decisions by layer-wise relevance …
Starting from the interpretability, this paper proposes the mechanism analysis idea of deep …

… We also designed rich and effective deep brain stimulation (DBS) control strategies for
epilepsy and gave a dynamic explanation of electrical stimulation to control absence epileptic …

: 本文驗證台指選擇權隱含波動率的淨買壓假說, 我們發現研究期間之台指選擇權隱含波動率呈現
負偏斜, 而此負偏斜導因於淨買壓, 且相依於選擇權契約的存續期間. 在控制台灣市場存在訊息…

… 30 3.9 SVM classifier. There are two different data and we need to find a decision boundary
to separate two classes of data and the margin between the two is maximal. . . . . . . . . . . . . . . . …