Learning to Diagnose with LSTM Recurrent Neural Networks

Abstract

Clinical medical data, especially in the intensive care unit (ICU), consist of multivariate time series of observations. For each patient visit (or episode), sensor data and lab test results are recorded in the patient's Electronic Health Record (EHR). While potentially containing a wealth of insights, the data is difficult to mine effectively, owing to varying length, irregular sampling and missing data. Recurrent Neural Networks (RNNs), particularly those using Long Short-Term Memory (LSTM) hidden units, are powerful and increasingly popular models for learning from sequence data. They effectively model varying length sequences and capture long range dependencies. We present the first study to empirically evaluate the ability of LSTMs to recognize patterns in multivariate time series of clinical measurements. Specifically, we consider multilabel classification of diagnoses, training a model to classify 128 diagnoses given 13 frequently but irregularly sampled clinical measurements. First, we establish the effectiveness of a simple LSTM network for modeling clinical data. Then we demonstrate a straightforward and effective training strategy in which we replicate targets at each sequence step. Trained only on raw time series, our models outperform several strong baselines, including a multilayer perceptron trained on hand-engineered features.

References in corpus (4)

• Sequence to Sequence Learning with Neural Networks
• Recurrent Neural Network Regularization
• Beyond Short Snippets: Deep Networks for Video Classification
• Bach in 2014: Music Composition with Recurrent Neural Network

Cited by in corpus (126)

• RetainVis: Visual Analytics with Interpretable and Interactive Recurrent Neural Networks on Electronic Medical Records
• A Multi-Horizon Quantile Recurrent Forecaster
• Temporal Attention augmented Bilinear Network for Financial Time-Series Data Analysis
• Recurrent Neural Networks for Multivariate Time Series with Missing Values
• GluonCV and GluonNLP: Deep Learning in Computer Vision and Natural Language Processing
• Patient2Vec: A Personalized Interpretable Deep Representation of the Longitudinal Electronic Health Record
• An Empirical Evaluation of Deep Learning for ICD-9 Code Assignment using MIMIC-III Clinical Notes
• Reinforcement Learning in Healthcare: A Survey
• Distilling Knowledge from Deep Networks with Applications to Healthcare Domain
• SLEEPNET: Automated Sleep Staging System via Deep Learning
• GRAM: Graph-based Attention Model for Healthcare Representation Learning
• ISeeU: Visually interpretable deep learning for mortality prediction inside the ICU
• Exploring Interpretable LSTM Neural Networks over Multi-Variable Data
• Clinical Intervention Prediction and Understanding using Deep Networks
• An Improved Multi-Output Gaussian Process RNN with Real-Time Validation for Early Sepsis Detection
• Improving Hospital Mortality Prediction with Medical Named Entities and Multimodal Learning
• Benchmark of Deep Learning Models on Large Healthcare MIMIC Datasets
• Multi-Label Classification of Patient Notes a Case Study on ICD Code Assignment
• Neural Document Embeddings for Intensive Care Patient Mortality Prediction
• Deep EHR: Chronic Disease Prediction Using Medical Notes
• Short term prediction of demand for ride hailing services: A deep learning approach
• Attend and Diagnose: Clinical Time Series Analysis using Attention Models
• Interpreting a Recurrent Neural Network's Predictions of ICU Mortality Risk
• Multi-task Prediction of Disease Onsets from Longitudinal Lab Tests
• Checkpoint Ensembles: Ensemble Methods from a Single Training Process
• Feature Engineering and Forecasting via Derivative-free Optimization and Ensemble of Sequence-to-sequence Networks with Applications in Renewable Energy
• R2N2: Residual Recurrent Neural Networks for Multivariate Time Series Forecasting
• Anomaly Detection using Deep Autoencoders for in-situ Wastewater Systems Monitoring Data
• Modeling Long- and Short-Term Temporal Patterns with Deep Neural Networks
• Combining LSTM and Latent Topic Modeling for Mortality Prediction
• RAIM: Recurrent Attentive and Intensive Model of Multimodal Patient Monitoring Data
• Improving Long-Horizon Forecasts with Expectation-Biased LSTM Networks
• The Use of Autoencoders for Discovering Patient Phenotypes
• DeepHealth: Review and challenges of artificial intelligence in health informatics
• On Generalization Bounds of a Family of Recurrent Neural Networks
• GAMENet: Graph Augmented MEmory Networks for Recommending Medication Combination
• Representation Learning for Electronic Health Records
• Pre-training of Graph Augmented Transformers for Medication Recommendation
• Transfer Learning for Clinical Time Series Analysis using Recurrent Neural Networks
• Semi-supervised Rare Disease Detection Using Generative Adversarial Network
• Deep reinforcement learning for time series: playing idealized trading games
• Identifiable Phenotyping using Constrained Non-Negative Matrix Factorization
• Predicting Acute Kidney Injury at Hospital Re-entry Using High-dimensional Electronic Health Record Data
• Multi-variable LSTM neural network for autoregressive exogenous model
• Bayesian LSTMs in medicine
• Not to Cry Wolf: Distantly Supervised Multitask Learning in Critical Care
• Techniques for visualizing LSTMs applied to electrocardiograms
• Rare Disease Detection by Sequence Modeling with Generative Adversarial Networks
• Accelerating High-Strain Continuum-Scale Brittle Fracture Simulations with Machine Learning
• Improved Hierarchical Patient Classification with Language Model Pretraining over Clinical Notes
• Effective Representations of Clinical Notes
• A Self-Correcting Deep Learning Approach to Predict Acute Conditions in Critical Care
• Learning to Identify Patients at Risk of Uncontrolled Hypertension Using Electronic Health Records Data
• Medical Diagnosis From Laboratory Tests by Combining Generative and Discriminative Learning
• Hybrid Gradient Boosting Trees and Neural Networks for Forecasting Operating Room Data
• Multimodal deep learning for short-term stock volatility prediction
• Multimodal Machine Learning for Automated ICD Coding
• Uncertainty-Aware Lookahead Factor Models for Quantitative Investing
• Identify Susceptible Locations in Medical Records via Adversarial Attacks on Deep Predictive Models
• Deep Representation for Patient Visits from Electronic Health Records
• Autowarp: Learning a Warping Distance from Unlabeled Time Series Using Sequence Autoencoders
• Explainable Goal-Driven Agents and Robots -- A Comprehensive Review
• Multi-modality Sensor Data Classification with Selective Attention
• Intra-day Activity Better Predicts Chronic Conditions
• Heterogeneous Similarity Graph Neural Network on Electronic Health Records
• Making Good on LSTMs' Unfulfilled Promise
• Explainable Multivariate Time Series Classification: A Deep Neural Network Which Learns To Attend To Important Variables As Well As Informative Time Intervals
• Towards Better Long-range Time Series Forecasting using Generative Adversarial Networks
• Neural Skill Transfer from Supervised Language Tasks to Reading Comprehension
• Developing Knowledge-enhanced Chronic Disease Risk Prediction Models from Regional EHR Repositories
• Enhancing Model Interpretability and Accuracy for Disease Progression Prediction via Phenotype-Based Patient Similarity Learning
• ICD Coding from Clinical Text Using Multi-Filter Residual Convolutional Neural Network
• A Probabilistic Disease Progression Model for Predicting Future Clinical Outcome
• HOLMES: Health OnLine Model Ensemble Serving for Deep Learning Models in Intensive Care Units
• An attention model to analyse the risk of agitation and urinary tract infections in people with dementia
• Learning the Graphical Structure of Electronic Health Records with Graph Convolutional Transformer
• Disease phenotyping using deep learning: A diabetes case study
• A Tree Architecture of LSTM Networks for Sequential Regression with Missing Data
• Prediction of the onset of cardiovascular diseases from electronic health records using multi-task gated recurrent units
• Predicting Length of Stay in the Intensive Care Unit with Temporal Pointwise Convolutional Networks
• Predicting Diabetes Disease Evolution Using Financial Records and Recurrent Neural Networks
• MUFASA: Multimodal Fusion Architecture Search for Electronic Health Records
• Demand Forecasting for Platelet Usage: from Univariate Time Series to Multivariate Models
• Multi-Task Prediction of Clinical Outcomes in the Intensive Care Unit using Flexible Multimodal Transformers
• Machine Learning with Electronic Health Records is vulnerable to Backdoor Trigger Attacks
• Predicting Discharge Medications at Admission Time Based on Deep Learning
• Parkinson's Disease Diagnosis Using Deep Learning
• Bayesian Recurrent Framework for Missing Data Imputation and Prediction with Clinical Time Series
• Tensor Graph Convolutional Networks for Text Classification
• The Impact of Extraneous Variables on the Performance of Recurrent Neural Network Models in Clinical Tasks
• The Effectiveness of Multitask Learning for Phenotyping with Electronic Health Records Data
• Natural Language Processing for EHR-Based Computational Phenotyping
• DeepTag: inferring all-cause diagnoses from clinical notes in under-resourced medical domain
• Long-term Blood Pressure Prediction with Deep Recurrent Neural Networks
• Position-based Content Attention for Time Series Forecasting with Sequence-to-sequence RNNs
• Long-range Prediction of Vital Signs Using Generative Boosting via LSTM Networks
• Graph Spectral Embedding for Parsimonious Transmission of Multivariate Time Series
• Phonocardiographic Sensing using Deep Learning for Abnormal Heartbeat Detection
• LSTM recurrent neural network assisted aircraft stall prediction for enhanced situational awareness
• Intelligent EHRs: Predicting Procedure Codes From Diagnosis Codes
• UNITE: Uncertainty-based Health Risk Prediction Leveraging Multi-sourced Data
• Co-Morbidity Exploration on Wearables Activity Data Using Unsupervised Pre-training and Multi-Task Learning
• Automatic Device Classification from Network Traffic Streams of Internet of Things
• Supervised multi-specialist topic model with applications on large-scale electronic health record data
• A latent topic model for mining heterogenous non-randomly missing electronic health records data
• GraSSNet: Graph Soft Sensing Neural Networks
• Predicting pregnancy using large-scale data from a women's health tracking mobile application
• Improving Generalization of Sequence Encoder-Decoder Networks for Inverse Imaging of Cardiac Transmembrane Potential
• Continual Learning Augmented Investment Decisions
• Unsupervised Learning for Computational Phenotyping
• FORECAST-CLSTM: A New Convolutional LSTM Network for Cloudage Nowcasting
• MEETING BOT: Reinforcement Learning for Dialogue Based Meeting Scheduling
• Interpretable Structure-aware Document Encoders with Hierarchical Attention
• Prediction of MRI Hardware Failures based on Image Features using Time Series Classification
• Modelling the Neuroanatomical Progression of Alzheimer's Disease and Posterior Cortical Atrophy
• Leveraging Disease Progression Learning for Medical Image Recognition
• Teaching a Machine to Diagnose a Heart Disease; Beginning from digitizing scanned ECGs to detecting the Brugada Syndrome (BrS)
• Longitudinal Pooling & Consistency Regularization to Model Disease Progression from MRIs
• Longitudinal Self-Supervised Learning
• Medical data wrangling with sequential variational autoencoders
• Towards Generating Real-World Time Series Data
• Time-Guided High-Order Attention Model of Longitudinal Heterogeneous Healthcare Data
• Hardware realisation of nonlinear dynamical systems for and from biology
• Maintaining The Humanity of Our Models
• Using Clinical Drug Representations for Improving Mortality and Length of Stay Predictions
• Improving Clinical Outcome Predictions Using Convolution over Medical Entities with Multimodal Learning