0.25
0.5
0.75
1.25
1.5
1.75
2
Synthetic maps for navigating high-dimensional data spaces
Published on May 16, 2019165 Views
Related categories
Chapter list
Synthetic maps for navigating high-dimensional data spaces00:00
Graduate school00:09
cecam01:07
Complex data landscapes are everywhere01:59
How can I get a low-dimensional map from my data? - 104:04
How can I get a low-dimensional map from my data? - 204:28
Problems05:17
A critical difficulty: projection (without choosing the collective variables)05:35
What happens if the manifold containing the data is curved?06:42
What now? Approximately one dimensional, but can we map this to a line?07:39
Real-world data:08:18
… still, let’s do it!!!!!!!!!!10:27
Learn from Marshallese sailors11:59
Our perspective on data landscapes14:17
The topography of a data landscape15:40
Building a topography of a data landscape - 116:50
Building a topography of a data landscape - 218:49
Estimate the intrinsic dimension of the data set18:58
The intrinsic dimension22:21
Building a topography of a data landscape23:21
The k-nearest neighbor estimator23:28
The key problem: highly non-uniform densities23:31
Adaptive density estimate24:33
Obtaining a position dependent k25:39
Two different hypothesis - 125:40
Two different hypothesis - 225:44
Two different hypothesis - 325:47
Benchmarks on realistic densities26:41
Building a topography of a data landscape27:38
Finding the density peaks - 127:54
Finding the density peaks - 229:53
Finding the density peaks - 330:28
Finding the density peaks - 430:37
No optimization required…30:46
The clustering approach at work - 131:58
The clustering approach at work - 232:03
The clustering approach at work - 332:05
Building a topography of a data landscape32:14
Clustering a MD trajectory - 132:17
Clustering a MD trajectory - 235:13
Clustering a MD trajectory - 335:27
Clustering a MD trajectory - 436:09
Clustering a MD trajectory - 536:11
Density Peak clustering + MSM - 136:15
Density Peak clustering + MSM - 236:36
Density Peak clustering + MSM - 337:14
Folding of a 32 -residue protein (Villin headpiece) - 137:21
Folding of a 32 -residue protein (Villin headpiece) - 237:26
Folding of a 32 -residue protein (Villin headpiece) - 339:07
The topography of a data landscape41:26
Automatic recognition of protein families - 141:29
Automatic recognition of protein families - 243:39
Automatic recognition of protein families - 344:18
Analysis of a fMRI experiment (D. Amati, M. Maieron, F. Pizzagalli) - 145:41
Analysis of a fMRI experiment (D. Amati, M. Maieron, F. Pizzagalli) - 249:30
Analysis of a fMRI experiment (D. Amati, M. Maieron, F. Pizzagalli) - 349:32
Conclusions50:57