UDAMA: Unsupervised Domain Adaptation through Multi-discriminator Adversarial Training with Noisy Labels Improves Cardio-fitness Prediction

Yu Wu, Dimitris Spathis, Hong Jia, Ignacio Perez-Pozuelo, Tomas I. Gonzales, Soren Brage, Nicholas Wareham, Cecilia Mascolo
Proceedings of the 8th Machine Learning for Healthcare Conference, PMLR 219:863-883, 2023.

Abstract

Deep learning models have shown great promise in various healthcare monitoring applications. However, most healthcare datasets with high-quality (gold-standard) labels are small-scale, as directly collecting ground truth is often costly and time-consuming. As a result, models developed and validated on small-scale datasets often suffer from overfitting and do not generalize well to unseen scenarios. At the same time, large amounts of imprecise (silver-standard) labeled data, annotated by approximate methods with the help of modern wearables and in the absence of ground truth validation, are starting to emerge. However, due to measurement differences, this data displays significant label distribution shifts, which motivates the use of domain adaptation. To this end, we introduce UDAMA, a method with two key components: Unsupervised Domain Adaptation and Multi-discriminator Adversarial Training, where we pre-train on the silver-standard data and employ adversarial adaptation with the gold-standard data along with two domain discriminators. In particular, we showcase the practical potential of UDAMA by applying it to Cardio-respiratory fitness (CRF) prediction. CRF is a crucial determinant of metabolic disease and mortality, and it presents labels with various levels of noise (gold- and silver-standard), making it challenging to establish an accurate prediction model. Our results show promising performance by alleviating distribution shifts in various label shift settings. Additionally, by using data from two free-living cohort studies (Fenland and BBVS), we show that UDAMA consistently outperforms up to 12% compared to competitive transfer learning and state-of-the-art domain adaptation models, paving the way for leveraging noisy labeled data to improve fitness estimation at scale.

Cite this Paper

BibTeX
@InProceedings{pmlr-v219-wu23a, title = {UDAMA: Unsupervised Domain Adaptation through Multi-discriminator Adversarial Training with Noisy Labels Improves Cardio-fitness Prediction}, author = {Wu, Yu and Spathis, Dimitris and Jia, Hong and Perez-Pozuelo, Ignacio and Gonzales, Tomas I. and Brage, Soren and Wareham, Nicholas and Mascolo, Cecilia}, booktitle = {Proceedings of the 8th Machine Learning for Healthcare Conference}, pages = {863--883}, year = {2023}, editor = {Deshpande, Kaivalya and Fiterau, Madalina and Joshi, Shalmali and Lipton, Zachary and Ranganath, Rajesh and Urteaga, Iñigo and Yeung, Serene}, volume = {219}, series = {Proceedings of Machine Learning Research}, month = {11--12 Aug}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v219/wu23a/wu23a.pdf}, url = {https://proceedings.mlr.press/v219/wu23a.html}, abstract = {Deep learning models have shown great promise in various healthcare monitoring applications. However, most healthcare datasets with high-quality (gold-standard) labels are small-scale, as directly collecting ground truth is often costly and time-consuming. As a result, models developed and validated on small-scale datasets often suffer from overfitting and do not generalize well to unseen scenarios. At the same time, large amounts of imprecise (silver-standard) labeled data, annotated by approximate methods with the help of modern wearables and in the absence of ground truth validation, are starting to emerge. However, due to measurement differences, this data displays significant label distribution shifts, which motivates the use of domain adaptation. To this end, we introduce UDAMA, a method with two key components: Unsupervised Domain Adaptation and Multi-discriminator Adversarial Training, where we pre-train on the silver-standard data and employ adversarial adaptation with the gold-standard data along with two domain discriminators. In particular, we showcase the practical potential of UDAMA by applying it to Cardio-respiratory fitness (CRF) prediction. CRF is a crucial determinant of metabolic disease and mortality, and it presents labels with various levels of noise (gold- and silver-standard), making it challenging to establish an accurate prediction model. Our results show promising performance by alleviating distribution shifts in various label shift settings. Additionally, by using data from two free-living cohort studies (Fenland and BBVS), we show that UDAMA consistently outperforms up to 12% compared to competitive transfer learning and state-of-the-art domain adaptation models, paving the way for leveraging noisy labeled data to improve fitness estimation at scale.} }
Endnote
%0 Conference Paper %T UDAMA: Unsupervised Domain Adaptation through Multi-discriminator Adversarial Training with Noisy Labels Improves Cardio-fitness Prediction %A Yu Wu %A Dimitris Spathis %A Hong Jia %A Ignacio Perez-Pozuelo %A Tomas I. Gonzales %A Soren Brage %A Nicholas Wareham %A Cecilia Mascolo %B Proceedings of the 8th Machine Learning for Healthcare Conference %C Proceedings of Machine Learning Research %D 2023 %E Kaivalya Deshpande %E Madalina Fiterau %E Shalmali Joshi %E Zachary Lipton %E Rajesh Ranganath %E Iñigo Urteaga %E Serene Yeung %F pmlr-v219-wu23a %I PMLR %P 863--883 %U https://proceedings.mlr.press/v219/wu23a.html %V 219 %X Deep learning models have shown great promise in various healthcare monitoring applications. However, most healthcare datasets with high-quality (gold-standard) labels are small-scale, as directly collecting ground truth is often costly and time-consuming. As a result, models developed and validated on small-scale datasets often suffer from overfitting and do not generalize well to unseen scenarios. At the same time, large amounts of imprecise (silver-standard) labeled data, annotated by approximate methods with the help of modern wearables and in the absence of ground truth validation, are starting to emerge. However, due to measurement differences, this data displays significant label distribution shifts, which motivates the use of domain adaptation. To this end, we introduce UDAMA, a method with two key components: Unsupervised Domain Adaptation and Multi-discriminator Adversarial Training, where we pre-train on the silver-standard data and employ adversarial adaptation with the gold-standard data along with two domain discriminators. In particular, we showcase the practical potential of UDAMA by applying it to Cardio-respiratory fitness (CRF) prediction. CRF is a crucial determinant of metabolic disease and mortality, and it presents labels with various levels of noise (gold- and silver-standard), making it challenging to establish an accurate prediction model. Our results show promising performance by alleviating distribution shifts in various label shift settings. Additionally, by using data from two free-living cohort studies (Fenland and BBVS), we show that UDAMA consistently outperforms up to 12% compared to competitive transfer learning and state-of-the-art domain adaptation models, paving the way for leveraging noisy labeled data to improve fitness estimation at scale.
APA
Wu, Y., Spathis, D., Jia, H., Perez-Pozuelo, I., Gonzales, T.I., Brage, S., Wareham, N. & Mascolo, C.. (2023). UDAMA: Unsupervised Domain Adaptation through Multi-discriminator Adversarial Training with Noisy Labels Improves Cardio-fitness Prediction. Proceedings of the 8th Machine Learning for Healthcare Conference, in Proceedings of Machine Learning Research 219:863-883 Available from https://proceedings.mlr.press/v219/wu23a.html.

Related Material