FLARe: Forecasting by Learning Anticipated Representations

Surya Teja Devarakonda, Joie Yeahuay Wu, Yi Ren Fung, Madalina Fiterau
Proceedings of the 4th Machine Learning for Healthcare Conference, PMLR 106:53-65, 2019.

Abstract

Computational models that forecast the progression of Alzheimer’s disease at the patient level are extremely useful tools for identifying high risk cohorts for early intervention and treatment planning. The state-of-the-art work in this area proposes models that forecast by using latent representations extracted from the longitudinal data across multiple modalities, including volumetric information extracted from medical scans and demographic info. These models incorporate the time horizon, which is the amount of time between the last recorded visit and the future visit, by directly concatenating a representation of it to the latent data representation. In this paper, we present a model which generates a sequence of latent representations of the patient status across the time horizon, providing more informative modeling of the temporal relationships between the patient’s history and future visits. Our proposed model outperforms the baseline in terms of forecasting accuracy and F1 score.

Cite this Paper

BibTeX
@InProceedings{pmlr-v106-devarakonda19a, title = {FLARe: Forecasting by Learning Anticipated Representations}, author = {Devarakonda, Surya Teja and Wu, Joie Yeahuay and Fung, Yi Ren and Fiterau, Madalina}, booktitle = {Proceedings of the 4th Machine Learning for Healthcare Conference}, pages = {53--65}, year = {2019}, editor = {Doshi-Velez, Finale and Fackler, Jim and Jung, Ken and Kale, David and Ranganath, Rajesh and Wallace, Byron and Wiens, Jenna}, volume = {106}, series = {Proceedings of Machine Learning Research}, month = {09--10 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v106/devarakonda19a/devarakonda19a.pdf}, url = {https://proceedings.mlr.press/v106/devarakonda19a.html}, abstract = {Computational models that forecast the progression of Alzheimer’s disease at the patient level are extremely useful tools for identifying high risk cohorts for early intervention and treatment planning. The state-of-the-art work in this area proposes models that forecast by using latent representations extracted from the longitudinal data across multiple modalities, including volumetric information extracted from medical scans and demographic info. These models incorporate the time horizon, which is the amount of time between the last recorded visit and the future visit, by directly concatenating a representation of it to the latent data representation. In this paper, we present a model which generates a sequence of latent representations of the patient status across the time horizon, providing more informative modeling of the temporal relationships between the patient’s history and future visits. Our proposed model outperforms the baseline in terms of forecasting accuracy and F1 score.} }
Endnote
%0 Conference Paper %T FLARe: Forecasting by Learning Anticipated Representations %A Surya Teja Devarakonda %A Joie Yeahuay Wu %A Yi Ren Fung %A Madalina Fiterau %B Proceedings of the 4th Machine Learning for Healthcare Conference %C Proceedings of Machine Learning Research %D 2019 %E Finale Doshi-Velez %E Jim Fackler %E Ken Jung %E David Kale %E Rajesh Ranganath %E Byron Wallace %E Jenna Wiens %F pmlr-v106-devarakonda19a %I PMLR %P 53--65 %U https://proceedings.mlr.press/v106/devarakonda19a.html %V 106 %X Computational models that forecast the progression of Alzheimer’s disease at the patient level are extremely useful tools for identifying high risk cohorts for early intervention and treatment planning. The state-of-the-art work in this area proposes models that forecast by using latent representations extracted from the longitudinal data across multiple modalities, including volumetric information extracted from medical scans and demographic info. These models incorporate the time horizon, which is the amount of time between the last recorded visit and the future visit, by directly concatenating a representation of it to the latent data representation. In this paper, we present a model which generates a sequence of latent representations of the patient status across the time horizon, providing more informative modeling of the temporal relationships between the patient’s history and future visits. Our proposed model outperforms the baseline in terms of forecasting accuracy and F1 score.
APA
Devarakonda, S.T., Wu, J.Y., Fung, Y.R. & Fiterau, M.. (2019). FLARe: Forecasting by Learning Anticipated Representations. Proceedings of the 4th Machine Learning for Healthcare Conference, in Proceedings of Machine Learning Research 106:53-65 Available from https://proceedings.mlr.press/v106/devarakonda19a.html.

Related Material