Patient Subtyping via Time-Aware LSTM Networks

@article{Baytas2017PatientSV,
  title={Patient Subtyping via Time-Aware LSTM Networks},
  author={Inci M. Baytas and Cao Xiao and Xi Sheryl Zhang and Fei Wang and Anil K. Jain and Jiayu Zhou},
  journal={Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining},
  year={2017}
}
  • Inci M. Baytas, Cao Xiao, +3 authors Jiayu Zhou
  • Published 2017
  • Computer Science
  • Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining
In the study of various diseases, heterogeneity among patients usually leads to different progression patterns and may require different types of therapeutic intervention. [...] Key Method The LSTM units are designed to handle data with constant elapsed times between consecutive elements of a sequence. Given that time lapse between successive elements in patient records can vary from days to months, the design of traditional LSTM may lead to suboptimal performance. In this paper, we propose a novel LSTM unit…Expand
A Time-Sensitive Hybrid Learning Model for Patient Subgrouping
TLDR
A time-sensitive hybrid learning model is proposed which utilizes the time intervals to control the decaying degree of history memories in patients’ pairwise representations and can effectively improve the validity of patient subgrouping. Expand
Temporal Phenotyping using Deep Predictive Clustering of Disease Progression
TLDR
A deep learning approach for clustering time-series data, where each cluster comprises patients who share similar future outcomes of interest (e.g., adverse events, the onset of comorbidities). Expand
Identifying Sepsis Subphenotypes via Time-Aware Multi-Modal Auto-Encoder
TLDR
A new sepsis subtyping framework is proposed which introduces time-aware attention mechanism and incorporates multi-modal inputs to impute missing values, a dynamic time wrapping (DTW) method to measure patients' temporal similarity based on the imputed EHR data, and a weighted k-means algorithm to cluster patients. Expand
Identifying Sepsis Subphenotypes via Time-Aware Multi-Modal Auto-Encoder
TLDR
A new sepsis subtyping framework is proposed which introduces time-aware attention mechanism and incorporates multi-modal inputs to impute missing values, a dynamic time wrapping (DTW) method to measure patients' temporal similarity based on the imputed EHR data, and a weighted k-means algorithm to cluster patients. Expand
Predicting Treatment Initiation from Clinical Time Series Data via Graph-Augmented Time-Sensitive Model
TLDR
Results show that relational similarity can improve prediction over multiple baselines, for example a 5% incremental over long-short term memory baseline in terms of area under precision-recall curve. Expand
Learning Temporal Relevance in Longitudinal Medical Notes
TLDR
This work proposes a novel attention mechanism composed of dual-attention blocks based on a rich diversity of time representations that outperforms state-of-the-art methods on six medical tasks using data sets from the publicly-available MIMIC III database from the Beth Israel Deaconess Medical Center. Expand
HiTANet: Hierarchical Time-Aware Attention Networks for Risk Prediction on Electronic Health Records
TLDR
A new hierarchical time-aware attention network, named HiTANet, which imitates the decision making process of doctors in risk prediction, which models time information in local and global stages and achieves over 7% in terms of F1 score on all datasets. Expand
DATA-GRU: Dual-Attention Time-Aware Gated Recurrent Unit for Irregular Multivariate Time Series
TLDR
A novel end-to-end Dual-Attention Time-Aware Gated Recurrent Unit (DATA-GRU) is proposed for IMTS to predict the mortality risk of patients and can significantly outperform state-of-the-art methods and provide meaningful clinical interpretation. Expand
Temporal Cascade and Structural Modelling of EHRs for Granular Readmission Prediction
TLDR
A novel model, MEDCAS, is proposed that combines the strengths of RNN-based models and point processes by integrating point processes in modelling visit types and time gaps into an attention-based sequence-tosequence learning model, which is able to capture the temporal cascade relationships. Expand
Time-aware Adversarial Networks for Adapting Disease Progression Modeling
TLDR
A domain adaptation (DA) framework named Time-aware Adversarial Networks is proposed, which combines two temporal LSTM-based recurrent components, a label predictor and a domain classifier, to extract an invariant feature representation across different patient groups through an adversarial learning process. Expand
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 37 REFERENCES
Learning to Diagnose with LSTM Recurrent Neural Networks
TLDR
This first study to empirically evaluate the ability of LSTMs to recognize patterns in multivariate time series of clinical measurements considers multilabel classification of diagnoses, and establishes the effectiveness of a simple LSTM network for modeling clinical data. Expand
Predicting Clinical Events by Combining Static and Dynamic Information Using Recurrent Neural Networks
TLDR
This work works with a database collected in the Charité Hospital in Berlin that contains complete information concerning patients that underwent a kidney transplantation, and develops an approach based on RNNs, specifically designed for the clinical domain, that combines static and dynamic information in order to predict future events. Expand
An RNN Architecture with Dynamic Temporal Matching for Personalized Predictions of Parkinson's Disease
TLDR
A deep model is proposed that directly learns patient similarity from longitudinal and multi-modal patient records with an Recurrent Neural Network (RNN) architecture, which learns the similarity between two longitudinal patient record sequences through dynamically matching temporal patterns in patient sequences. Expand
Deep Computational Phenotyping
TLDR
This work proposes two novel modifications to standard neural net training that address challenges and exploit properties that are peculiar, if not exclusive, to medical data, and examines a general framework for using prior knowledge to regularize parameters in the topmost layers. Expand
DeepCare: A Deep Dynamic Memory Model for Predictive Medicine
TLDR
The efficacy of DeepCare for disease progression modeling and readmission prediction in diabetes, a chronic disease with large economic burden, is demonstrated and the results show improved modeling and risk prediction accuracy. Expand
From micro to macro: data driven phenotyping by densification of longitudinal electronic medical records
TLDR
A data driven phenotyping framework called Pacifier (PAtient reCord densIFIER), where the longitudinal EMR data of each patient is interpreted as a sparse matrix with a feature dimension and a time dimension, and more robust patient phenotypes are derived by exploring the latent structure of those matrices. Expand
Recurrent Neural Networks for Multivariate Time Series with Missing Values
TLDR
Novel deep learning models are developed based on Gated Recurrent Unit, a state-of-the-art recurrent neural network that takes two representations of missing patterns, i.e., masking and time interval, and effectively incorporates them into a deep model architecture so that it not only captures the long-term temporal dependencies in time series, but also utilizes the missing patterns to achieve better prediction results. Expand
Multi-layer Representation Learning for Medical Concepts
TLDR
This work proposes Med2Vec, which not only learns the representations for both medical codes and visits from large EHR datasets with over million visits, but also allows us to interpret the learned representations confirmed positively by clinical experts. Expand
Long Short-Term Memory
TLDR
A novel, efficient, gradient based method called long short-term memory (LSTM) is introduced, which can learn to bridge minimal time lags in excess of 1000 discrete-time steps by enforcing constant error flow through constant error carousels within special units. Expand
A multi-task learning formulation for predicting disease progression
TLDR
This paper proposes a multi-task learning formulation for predicting the disease progression measured by the cognitive scores and selecting markers predictive of the progression and shows that most markers selected by the proposed algorithm are consistent with findings from existing cross-sectional studies. Expand
...
1
2
3
4
...