Bridging the gap

Simultaneous fine tuning for data re-balancing

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

There are many real-world classification problems wherein the issue of data imbalance (the case when a data set contains substantially more samples for one/many classes than the rest) is unavoidable. While under-sampling the problematic classes is a common solution, this is not a compelling option when the large data class is itself diverse and/or the limited data class is especially small. We suggest a strategy based on recent work concerning limited data problems which utilizes a supplemental set of images with similar properties to the limited data class to aid in the training of a neural network. We show results for our model against other typical methods on a real-world synthetic aperture sonar data set. Code can be found at github.com/JohnMcKay/dataImbalance.

Original languageEnglish (US)
Title of host publication2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages7062-7065
Number of pages4
ISBN (Electronic)9781538671504
DOIs
StatePublished - Oct 31 2018
Event38th Annual IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Valencia, Spain
Duration: Jul 22 2018Jul 27 2018

Publication series

NameInternational Geoscience and Remote Sensing Symposium (IGARSS)
Volume2018-July

Conference

Conference38th Annual IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018
CountrySpain
CityValencia
Period7/22/187/27/18

Fingerprint

Synthetic aperture sonar
Tuning
Sampling
Neural networks
sonar
sampling

All Science Journal Classification (ASJC) codes

  • Computer Science Applications
  • Earth and Planetary Sciences(all)

Cite this

McKay, J., Gerg, I. D., & Monga, V. (2018). Bridging the gap: Simultaneous fine tuning for data re-balancing. In 2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings (pp. 7062-7065). [8518664] (International Geoscience and Remote Sensing Symposium (IGARSS); Vol. 2018-July). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IGARSS.2018.8518664
McKay, John ; Gerg, Isaac D. ; Monga, Vishal. / Bridging the gap : Simultaneous fine tuning for data re-balancing. 2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2018. pp. 7062-7065 (International Geoscience and Remote Sensing Symposium (IGARSS)).
@inproceedings{72de7a99bf9d46509c3d9210aa11355e,
title = "Bridging the gap: Simultaneous fine tuning for data re-balancing",
abstract = "There are many real-world classification problems wherein the issue of data imbalance (the case when a data set contains substantially more samples for one/many classes than the rest) is unavoidable. While under-sampling the problematic classes is a common solution, this is not a compelling option when the large data class is itself diverse and/or the limited data class is especially small. We suggest a strategy based on recent work concerning limited data problems which utilizes a supplemental set of images with similar properties to the limited data class to aid in the training of a neural network. We show results for our model against other typical methods on a real-world synthetic aperture sonar data set. Code can be found at github.com/JohnMcKay/dataImbalance.",
author = "John McKay and Gerg, {Isaac D.} and Vishal Monga",
year = "2018",
month = "10",
day = "31",
doi = "10.1109/IGARSS.2018.8518664",
language = "English (US)",
series = "International Geoscience and Remote Sensing Symposium (IGARSS)",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "7062--7065",
booktitle = "2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings",
address = "United States",

}

McKay, J, Gerg, ID & Monga, V 2018, Bridging the gap: Simultaneous fine tuning for data re-balancing. in 2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings., 8518664, International Geoscience and Remote Sensing Symposium (IGARSS), vol. 2018-July, Institute of Electrical and Electronics Engineers Inc., pp. 7062-7065, 38th Annual IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018, Valencia, Spain, 7/22/18. https://doi.org/10.1109/IGARSS.2018.8518664

Bridging the gap : Simultaneous fine tuning for data re-balancing. / McKay, John; Gerg, Isaac D.; Monga, Vishal.

2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2018. p. 7062-7065 8518664 (International Geoscience and Remote Sensing Symposium (IGARSS); Vol. 2018-July).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Bridging the gap

T2 - Simultaneous fine tuning for data re-balancing

AU - McKay, John

AU - Gerg, Isaac D.

AU - Monga, Vishal

PY - 2018/10/31

Y1 - 2018/10/31

N2 - There are many real-world classification problems wherein the issue of data imbalance (the case when a data set contains substantially more samples for one/many classes than the rest) is unavoidable. While under-sampling the problematic classes is a common solution, this is not a compelling option when the large data class is itself diverse and/or the limited data class is especially small. We suggest a strategy based on recent work concerning limited data problems which utilizes a supplemental set of images with similar properties to the limited data class to aid in the training of a neural network. We show results for our model against other typical methods on a real-world synthetic aperture sonar data set. Code can be found at github.com/JohnMcKay/dataImbalance.

AB - There are many real-world classification problems wherein the issue of data imbalance (the case when a data set contains substantially more samples for one/many classes than the rest) is unavoidable. While under-sampling the problematic classes is a common solution, this is not a compelling option when the large data class is itself diverse and/or the limited data class is especially small. We suggest a strategy based on recent work concerning limited data problems which utilizes a supplemental set of images with similar properties to the limited data class to aid in the training of a neural network. We show results for our model against other typical methods on a real-world synthetic aperture sonar data set. Code can be found at github.com/JohnMcKay/dataImbalance.

UR - http://www.scopus.com/inward/record.url?scp=85064183292&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85064183292&partnerID=8YFLogxK

U2 - 10.1109/IGARSS.2018.8518664

DO - 10.1109/IGARSS.2018.8518664

M3 - Conference contribution

T3 - International Geoscience and Remote Sensing Symposium (IGARSS)

SP - 7062

EP - 7065

BT - 2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

McKay J, Gerg ID, Monga V. Bridging the gap: Simultaneous fine tuning for data re-balancing. In 2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2018. p. 7062-7065. 8518664. (International Geoscience and Remote Sensing Symposium (IGARSS)). https://doi.org/10.1109/IGARSS.2018.8518664