Imperial College London
Portrait Picture

DR BERNHARD KAINZ

Faculty of Engineering, Department of Computing

Reader in Medical Image Computing
 
 
 
//

Contact

 

+44 (0)20 7594 8349b.kainz Website CV

 
 
//

Location

 

372Huxley BuildingSouth Kensington Campus

//

Summary

 

Publications

Citation

BibTex format

@article{Castro:2019,
author = {Castro, DC and Tan, J and Kainz, B and Konukoglu, E and Glocker, B},
journal = {Journal of Machine Learning Research},
pages = {1--29},
title = {Morpho-MNIST: quantitative assessment and diagnostics for representation learning},
url = {http://arxiv.org/abs/1809.10780v2},
volume = {20},
year = {2019}
}
Download

RIS format (EndNote, RefMan)

TY  - JOUR
AB  - Revealing latent structure in data is an active field of research, havingintroduced exciting technologies such as variational autoencoders andadversarial networks, and is essential to push machine learning towardsunsupervised knowledge discovery. However, a major challenge is the lack ofsuitable benchmarks for an objective and quantitative evaluation of learnedrepresentations. To address this issue we introduce Morpho-MNIST, a frameworkthat aims to answer: "to what extent has my model learned to represent specificfactors of variation in the data?" We extend the popular MNIST dataset byadding a morphometric analysis enabling quantitative comparison of trainedmodels, identification of the roles of latent variables, and characterisationof sample diversity. We further propose a set of quantifiable perturbations toassess the performance of unsupervised and supervised methods on challengingtasks such as outlier detection and domain adaptation. Data and code areavailable at https://github.com/dccastro/Morpho-MNIST.
AU  - Castro,DC
AU  - Tan,J
AU  - Kainz,B
AU  - Konukoglu,E
AU  - Glocker,B
EP  - 29
PY  - 2019///
SN  - 1532-4435
SP  - 1
TI  - Morpho-MNIST: quantitative assessment and diagnostics for representation learning
T2  - Journal of Machine Learning Research
UR  - http://arxiv.org/abs/1809.10780v2
UR  - http://www.jmlr.org/papers/volume20/19-033/19-033.pdf
UR  - http://hdl.handle.net/10044/1/80003
VL  - 20
ER  -
Download