Representation Learning for Early Sepsis Prediction

As part of the PhysioNet/Computing in Cardiology Challenge 2019, we propose a neural network called AEC-Net to early detect sepsis based on physiological data. AEC-Net consists of two main components: 1) an Auto Encoder for dimension reduction and feature extraction, and 2) a Fully Connected Neural Network (FCNN) taking the extracted features by the Auto Encoder as the input and generating prediction of sepsis as output. The losses of both the Auto Encoder and FCNN are minimized concurrently. This concurrent optimization helps AEC-Net to have a better generalization and the extracted features by Auto Encoder to be more relevant to the classification problem. Finally, we propose an ensemble method of AEC-Net, Random Forest and Gradient Boosting Decision Trees to achieve a better prediction.We train our proposed models using data from 40336 patients with 40 physiological features ranging from 8 to 336 hours. Our team Infolab USC evaluated Ensemble with the hidden full test set of the Physionet Challenge 2019, and achieved a Utility score of 0.284 and 24th place in the challenge.

[1]  K. Wood,et al.  Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock* , 2006, Critical care medicine.

[2]  Kurt Hornik,et al.  Approximation capabilities of multilayer feedforward networks , 1991, Neural Networks.

[3]  Ashish Sharma,et al.  Early Prediction of Sepsis from Clinical Data: the PhysioNet/Computing in Cardiology Challenge 2019 , 2019, 2019 Computing in Cardiology (CinC).

[4]  Erik Marchi,et al.  A novel approach for automatic acoustic novelty detection using a denoising autoencoder with bidirectional LSTM neural networks , 2015, 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[5]  Pascal Vincent,et al.  Representation Learning: A Review and New Perspectives , 2012, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[6]  Piotr Indyk,et al.  Approximate nearest neighbors: towards removing the curse of dimensionality , 1998, STOC '98.

[7]  Tie-Yan Liu,et al.  LightGBM: A Highly Efficient Gradient Boosting Decision Tree , 2017, NIPS.

[8]  Steven Horng,et al.  Creating an automated trigger for sepsis clinical decision support at emergency department triage using machine learning , 2017, PloS one.

[9]  W. Knaus,et al.  Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. 1992. , 2009, Chest.

[10]  The quick sequential organ failure assessment (qSOFA) identifies septic patients in the out‐of‐hospital setting , 2018, The American journal of emergency medicine.

[11]  Uli K. Chettipally,et al.  Multicentre validation of a sepsis prediction algorithm using only vital sign data in the emergency department, general ward and ICU , 2018, BMJ Open.

[12]  J. Vincent,et al.  The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure , 1996, Intensive Care Medicine.

[13]  Geoffrey E. Hinton,et al.  Speech recognition with deep recurrent neural networks , 2013, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing.

[14]  Hye Jin Kam,et al.  Learning representations for the early detection of sepsis with deep neural networks , 2017, Comput. Biol. Medicine.

[15]  C. Subbe,et al.  Validation of a modified Early Warning Score in medical admissions. , 2001, QJM : monthly journal of the Association of Physicians.

[16]  Shamim Nemati,et al.  Early Prediction of Sepsis From Clinical Data: The PhysioNet/Computing in Cardiology Challenge 2019 , 2019, 2019 Computing in Cardiology (CinC).

[17]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.