相关实验视频

Updated: Jan 20, 2026

Data Acquisition Protocol for Determining Embedded Sensitivity Functions

Data Acquisition Protocol for Determining Embedded Sensitivity Functions

Published on: April 20, 2016

可视化稳定性:对于t-SNE嵌入的灵敏度分析框架

Susanne Zabel¹, Philipp Hennig², Kay Nieselt¹

¹Institute for Bioinformatics and Medical Informatics, University of Tübingen, Tübingen, Germany.

Frontiers in bioinformatics

|January 19, 2026

概括

本研究引入了一个计算框架来增强t分布式静态邻居嵌入 (t-SNE) 可视化. 它为稳定性和特征影响增加了视觉线索,改善了生物数据的解释.

科学领域:

生物信息学是一种生物信息学.
计算生物学计算生物学
数据可视化数据可视化

背景情况:

t-分布式静态邻居嵌入 (t-SNE) 广泛用于可视化高维生物数据.
标准t-SNE图中缺乏关于可视化稳定性,特征影响和数据不确定性的信息.

研究的目的:

开发一个用于扩展t-SNE图表的计算框架,并提供嵌入稳定性的视觉线索.
在t-SNE可视化中识别有影响力的特征并量化位置不确定性.

主要方法:

使用隐式函数定理和自动微分来计算第一阶导数的雅可比式的灵敏度分析.
通过雅可比数来传播输入数据的不确定性,以量化嵌入点的位置不确定性.
用热图可视化特征影响和假设结果的位置不确定性.

主要成果:

该框架揭示了在t-SNE嵌入中具有影响力的输入特征和结构不稳定的区域.
在输入数据不确定性可用时,对每个嵌入点的位置不确定性进行概率定量和可视化.
成功应用于大量RNA-seq,蛋白质组学和单细胞转录组学数据集.

结论:

开发的框架提供了一种原则方法来评估t-SNE可视化的稳定性和可解释性.
通过将视觉模式与生物驱动因素联系起来,并揭示模两可的情况,使得科学结论更加严格和明智.
增强t-SNE在生物信息学中的实用性,用于分析复杂的生物数据集.

关键词:

数据洞察力数据洞察力错误传播的传播是错误的传播可以解释的机器学习这就是T-SNENE.不确定性是一种不确定性.视觉化的可视化

更多相关视频

Watershed Planning within a Quantitative Scenario Analysis Framework

Watershed Planning within a Quantitative Scenario Analysis Framework

Published on: July 24, 2016

Constructing and Visualizing Models using Mime-based Machine-learning Framework

Constructing and Visualizing Models using Mime-based Machine-learning Framework

Published on: July 22, 2025

相关概念视频

Data Acquisition Protocol for Determining Embedded Sensitivity Functions

Data Acquisition Protocol for Determining Embedded Sensitivity Functions

The data acquisition procedure for determining embedded sensitivity functions is described. Data is acquired and representative results are shown for a residential scale wind turbine...

Watershed Planning within a Quantitative Scenario Analysis Framework

Watershed Planning within a Quantitative Scenario Analysis Framework

There is a critical need for tools and methodologies capable of managing aquatic systems in the face of uncertain future conditions. We provide methods for conducting a targeted watershed assessment that enables resource managers to produce landscape-based cumulative effects models for use within a scenario analysis management...

Constructing and Visualizing Models using Mime-based Machine-learning Framework

Constructing and Visualizing Models using Mime-based Machine-learning Framework

Mime is a flexible computational framework to construct a machine learning-based integration model with elegant performance. Here, we provide a detailed step-by-step procedure for developing predictive models with high accuracy, leveraging complex datasets to identify critical genes associated with disease progression, patient outcomes, and therapeutic response.

Plasmid Stability Analysis with Open-Source Droplet Microfluidics

Plasmid Stability Analysis with Open-Source Droplet Microfluidics

An accessible, open-source microfluidic workflow is presented for the parallelized analysis of plasmid retention in bacteria. By employing fluorescence microscopy to quantify plasmid presence in single-cell microcolonies within gel microdroplets, this method provides a precise, accessible, and scalable alternative to traditional plate...

Exploring Infant Sensitivity to Visual Language using Eye Tracking and the Preferential Looking Paradigm

Exploring Infant Sensitivity to Visual Language using Eye Tracking and the Preferential Looking Paradigm

Eye tracking studies using a preferential looking paradigm can be used to study infants' emerging understanding of, and attention to, their external visual...

Testing Visual Sensitivity to the Speed and Direction of Motion in Lizards

Testing Visual Sensitivity to the Speed and Direction of Motion in Lizards

Testing visual sensitivity in lizards using an operant conditioning paradigm that employs video playback of random-dot kinematograms and computer-generated...