TY  - CONF
ID  - kauschke18mil
T1  - Leveraging Reproduction-Error Representations for Multi-Instance Classification
A1  - Kauschke, Sebastian
A1  - Mühlhäuser, Max
A1  - Fürnkranz, Johannes
TI  - Proceedings of the 21st International Conference on Discovery Science - DS'18
T3  - Lecture Notes on Artificial Intelligence
Y1  - 2018
VL  - 11198
SP  - 83
EP  - 95
PB  - Springer Nature Switzerland AG
CY  - Limassol, Cyprus
M2  - doi: https://doi.org/10.1007/978-3-030-01771-2_6
KW  - Bag Classification
KW  - Denoising Autoencoder
KW  - Multi-Instance Learning
KW  - Reproduction-Error Representation
N2  - Multi-instance learning deals with the problem of classifying
bags of instances, when only the labels of the bags are known for
learning, and the instances themselves have no labels. In this work, we
propose a method that trains autoencoders for the instances in each
class, and recodes each instance into a representation that captures the
reproduction error for this instance. The idea behind this approach is
that an autoencoder trained on only instances of a single class is unable
to reproduce examples from another class properly, which is then
reflected in the encoding. The transformed instances are then piped into
a propositional classifier that decides the latent instance label. In a second
classification layer, the bag label is decided based on the output of
the propositional classifier on all the instances in the bag. We show that
this reproduction-error encoding creates an advantage compared to the
classification of non-encoded data, and that further research into this
direction could be beneficial for the cause of multi-instance learning.
ER  -