Incorporating Prior Domain Knowledge into Deep Neural Networks

Nikhil Muralidhar, Mohammad Raihanul Islam, Manish Marwah, Anuj Karpatne, Naren Ramakrishnan

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

123 Scopus citations

Abstract

In recent years, the large amount of labeled data available has also helped tend research toward using minimal domain knowledge, e.g., in deep neural network research. However, in many situations, data is limited and of poor quality. Can domain knowledge be useful in such a setting? In this paper, we propose domain adapted neural networks (DANN) to explore how domain knowledge can be integrated into model training for deep networks. In particular, we incorporate loss terms for knowledge available as monotonicity constraints and approximation constraints. We evaluate our model on both synthetic data generated using the popular Bohachevsky function and a real-world dataset for predicting oxygen solubility in water. In both situations, we find that our DANN model outperforms its domain-agnostic counterpart yielding an overall mean performance improvement of 19.5% with a worst- and best-case performance improvement of 4% and 42.7%, respectively.

Original languageEnglish
Title of host publicationProceedings - 2018 IEEE International Conference on Big Data, Big Data 2018
EditorsNaoki Abe, Huan Liu, Calton Pu, Xiaohua Hu, Nesreen Ahmed, Mu Qiao, Yang Song, Donald Kossmann, Bing Liu, Kisung Lee, Jiliang Tang, Jingrui He, Jeffrey Saltz
Pages36-45
Number of pages10
ISBN (Electronic)9781538650356
DOIs
StatePublished - 2 Jul 2018
Event2018 IEEE International Conference on Big Data, Big Data 2018 - Seattle, United States
Duration: 10 Dec 201813 Dec 2018

Publication series

NameProceedings - 2018 IEEE International Conference on Big Data, Big Data 2018

Conference

Conference2018 IEEE International Conference on Big Data, Big Data 2018
Country/TerritoryUnited States
CitySeattle
Period10/12/1813/12/18

Keywords

  • Deep Learning
  • Domain Knowledge
  • Limited Training Data
  • Neural Networks
  • Noisy Data

Fingerprint

Dive into the research topics of 'Incorporating Prior Domain Knowledge into Deep Neural Networks'. Together they form a unique fingerprint.

Cite this