Current Location:home > Browse

1. chinaXiv:202110.00004 [pdf]

深度学习在低频非侵入式负荷分解中的应用与比较

Li, Yiming; Ju, Yuntao; Qu, Liao
Subjects: Computer Science >> Other Disciplines of Computer Science

非侵入式负荷分解能够充分解析用户用电数据,是分析评估用户柔性调控潜力的关键技术。鉴于深度学习方法在负荷分解的广泛应用,首先深入探讨了降噪自编码器、循环神经网络、卷积神经网络等主流深度学习网络结构应用在负荷分解问题时的工作机理,并对它们在负荷分解领域中的应用与发展进行了展望分析;之后,提出了基于不同维度的分解算法评价框架,并补充了评价过程中测试数据的选择规范;最后,利用该评价框架对主流的深度学习分解模型进行评价,并对模型代码进行了开源,评价结果证明所提框架更能综合地评价给定超参设置下的深度学习分解模型,并揭示模型性能关于网络结构、特征输入等因素的敏感性。

submitted time 2021-10-01 Hits2215Downloads149 Comment 0

2. chinaXiv:201605.00743 [pdf]

CRASP: CFP reconstitution across synaptic partners

Li, Yiming; Guo, Aike; Li, Hao; Li, Yiming; Guo, Aike;
Subjects: Biology >> Biophysics >> Biochemistry & Molecular Biology

Mapping the pattern of connectivity between neurons is widely regarded to be critical for understanding the nervous system. GRASP (GFP reconstitution across synaptic partners) has been used as a promising method for mapping neuronal connectivity, but is currently available in the green color only, limiting its potential applications. Here we demonstrate CRASP (CFP reconstitution across synaptic partners), a cyan colored version of GRASP. We validated the system in HEK 293T cells, and generated transgenic Drosophila lines to show that the system could reliably detect neuronal contacts in the brain. Furthermore, we showed that the CRASP signal could be selectively amplified using standard immunohistochemistry methods. The CRASP system adds to the toolkit available to researchers for mapping neuronal connectivity, and substantially expands the potential application of GRASP-like strategies. (C) 2015 Elsevier Inc. All rights reserved.

submitted time 2016-05-05 Hits1779Downloads974 Comment 0

  [1 Pages/ 2 Totals]