{"type":"ScholarlyArticle","title":"Temporal Belief Memory: Imputing Missing Data during RNN Training","authors":[{"name":"Yeo Jin Kim","slug":"yeo-jin-kim","given":"Yeo Jin","family":"Kim","affiliation":"North Carolina State University"},{"name":"Min Chi","slug":"min-chi","given":"Min","family":"Chi","affiliation":"North Carolina State University"}],"abstract":"We propose a bio-inspired approach named Temporal Belief Memory (TBM) for handling missing data with recurrent neural networks (RNNs). When modeling irregularly observed temporal sequences, conventional RNNs generally ignore the real-time intervals between consecutive observations. TBM is a missing value imputation method that considers the time continuity and captures latent missing patterns based on irregular real time intervals of the inputs. We evaluate our TBM approach with real-world electronic health records (EHRs) consisting of 52,919 visits and 4,224,567 events on a task of early prediction of septic shock. We compare TBM against multiple baselines including both domain experts' rules and the state-of-the-art missing data handling approach using both RNN and long-short term memory. The experimental results show that TBM outperforms all the competitive baseline approaches for the septic shock early prediction task.","doi":"10.24963/ijcai.2018/322","url":"https://ijcai.org/proceedings/2018/322","pdf":"https://ijcai.org/proceedings/2018/0322.pdf","markdown":"https://ijcai.org/proceedings/2018/322.md","bibtex":"https://ijcai.org/proceedings/2018/bibtex/322","volume":{"slug":"2018","title":"Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2018","url":"https://ijcai.org/proceedings/2018/","year":2018,"location":"Stockholm","dates":"13-19 July 2018","editors":"Jérôme Lang","isbn":"978-0-9992411-2-7","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2018"},"section":"Machine Learning","track":"Main track","number":322,"pages":{"first":2326,"last":2332},"keywords":["Multidisciplinary Topics and Applications: Multidisciplinary Topics and Applications","Machine Learning: Deep Learning","Machine Learning Applications: Applications of Supervised Learning","Machine Learning: Time-series;Data Streams"],"videos":[],"references":[{"position":1,"raw":"[Bengio and Gingras, 1995] Yoshua Bengio and Francois Gingras. Recurrent neural networks for missing or asyn- chronous data. In NIPS, pages 395-401, 1995.","title":"Recurrent neural networks for missing or asynchronous data","authors":"Gingras Bengio, Yoshua Bengio, Francois Gingras","year":"1995","venue":"NIPS","doi":null,"url":null,"internal_url":null},{"position":2,"raw":"[Che et al., 2016] Zhengping Che, Sanjay Purushotham, Kyunghyun Cho, David Sontag, and Yan Liu. Recurrent neural networks for multivariate time series with missing values. CoRR, abs/1606.01865, 2016.","title":"Recurrent neural networks for multivariate time series with missing values","authors":"Zhengping Che, Sanjay Purushotham, Kyunghyun Cho, David Sontag, Yan Liu","year":"2016","venue":null,"doi":null,"url":null,"internal_url":null},{"position":3,"raw":"[Cho et al., 2014] Kyunghyun Cho, Bart van Merrienboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. Learning phrase representations using RNN encoder-decoder for statistical machine translation. In EMNLP, page 1724-1734, 2014.","title":"Learning phrase representations using RNN encoder-decoder for statistical machine translation","authors":"Cho","year":"2014","venue":"EMNLP","doi":null,"url":null,"internal_url":null},{"position":4,"raw":"[Cismondi et al., 2013] Federico Cismondi, Andre Fialho, Susana Vieira, Shane Reti, João Sousa, and Stan Finkel- stein. Missing data in medical databases: Impute, delete or classify. AI in Medicine, 58(1), May 2013.","title":"Missing data in medical databases: Impute, delete or classify","authors":"Federico Cismondi, Andre Fialho, Susana Vieira, Shane Reti, João Sousa, Stan Finkelstein","year":"2013","venue":"AI in Medicine","doi":null,"url":null,"internal_url":null},{"position":5,"raw":"[Cohen and Cohen, 1975] Jacob Cohen and Patricia Cohen. Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences. John Wiley&Sons, New York, 1975.","title":"Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences","authors":"Cohen Cohen, Jacob Cohen, Patricia Cohen","year":"1975","venue":null,"doi":null,"url":null,"internal_url":null},{"position":6,"raw":"[Galimard et al., 2016] Jacques-Emmanuel Galimard, Sylvie Chevret, Camelia Protopopescu, and Matthieu Resche-Rigon. A multiple imputation approach for mnar mechanisms compatible with heckman's model. Statistics in medicine, 35, February 2016.","title":"A multiple imputation approach for mnar mechanisms compatible with heckman's model","authors":"Jacques-Emmanuel Galimard, Sylvie Chevret, Camelia Protopopescu, Matthieu Resche-Rigon","year":"2016","venue":"Statistics in medicine","doi":null,"url":null,"internal_url":null},{"position":7,"raw":"[García-Laencina et al., 2015] Pedro García-Laencina, Pe- dro Abreu, Miguel Abreu, and Noémia Afonoso. Missing data imputation on the 5-year survival prediction of breast cancer patients with unknown discrete values. Computers in Biology and Medicine, 59(1), April 2015.","title":"Missing data imputation on the 5-year survival prediction of breast cancer patients with unknown discrete values","authors":"Pedro García-Laencina, Pedro Abreu, Miguel Abreu, Noémia Afonoso","year":"2015","venue":"Computers in Biology and Medicine","doi":null,"url":null,"internal_url":null},{"position":8,"raw":"[Gerstner and Kistler, 2002] Wulfram Gerstner and Werner Kistler. Spiking neuron models: single neurons, popula- tions, plasticity. Cambridge University Press, Cambridge, 2002.","title":"Spiking neuron models: single neurons, populations, plasticity","authors":"Kistler Gerstner, Wulfram Gerstner, Werner Kistler","year":"2002","venue":null,"doi":null,"url":null,"internal_url":null},{"position":9,"raw":"[Graves, 2013] Alex Graves. Generating sequences with re- current neural networks. CoRR, abs/1308.0850, 2013.","title":"Generating sequences with recurrent neural networks","authors":"Alex Graves","year":"2013","venue":null,"doi":null,"url":null,"internal_url":null},{"position":10,"raw":"[Ho et al., 2014] Joyce Ho, Cheng Lee, and Joydeep Ghosh. Septic shock prediction for patients with missing data. Management Information Systems, 5(1), April 2014.","title":"Septic shock prediction for patients with missing data","authors":"Joyce Ho, Cheng Lee, Joydeep Ghosh","year":"2014","venue":"Management Information Systems","doi":null,"url":null,"internal_url":null},{"position":11,"raw":"[Hochreiter and Schmidhuber, 1997] Sepp Hochreiter and Jurgen Schmidhuber. Long short-term memory. Neural Computation, 9(8), December 1997.","title":"Long short-term memory","authors":"Schmidhuber Hochreiter, Sepp Hochreiter, Jurgen Schmidhuber","year":"1997","venue":"Neural Computation","doi":null,"url":null,"internal_url":null},{"position":12,"raw":"[Hodgkin and Huxley, 1952] Alan Hodgkin and Andrew Huxley. A quantitative description of membrane current and its application to conduction and excitation in nerve. The Journal of Physiology, 117(4), August 1952.","title":"Alan Hodgkin and Andrew Huxley. A quantitative description of membrane current and its application to conduction and excitation in nerve","authors":"Huxley Hodgkin","year":"1952","venue":"The Journal of Physiology","doi":null,"url":null,"internal_url":null},{"position":13,"raw":"[Hsieh and Tang, 2013] Hung-Yi Hsieh and Kea-Tiong Tang. Hardware friendly probabilistic spiking neural network with long-term and short-term plasticity. Neural Networks and Learning Systems, 24(12), December 2013.","title":"Hardware friendly probabilistic spiking neural network with long-term and short-term plasticity","authors":"Tang ; ] Hung-Yi Hsieh, Kea-Tiong Hsieh, Tang","year":"2013","venue":"Neural Networks and Learning Systems","doi":null,"url":null,"internal_url":null},{"position":14,"raw":"[Kadri et al., 2017] Sameer Kadri, Chanu Rhee, Jeffrey Strich, Megan Morales, Samuel Hohmann, Jonathan Men- chaca, Anthony Suffredini, Robert Danner, and Michael Klompas. Estimating ten-year trends in septic shock in- cidence and mortality in united states academic medical centers using clinical data. Chest, 2017.","title":"Estimating ten-year trends in septic shock incidence and mortality in united states academic medical centers using clinical data","authors":"Sameer Kadri, Chanu Rhee, Jeffrey Strich, Megan Morales, Samuel Hohmann, Jonathan Menchaca, Anthony Suffredini, Robert Danner, Michael Klompas","year":"2017","venue":"Chest","doi":null,"url":null,"internal_url":null},{"position":15,"raw":"[Kim et al., 2017] Han-Gyu Kim, Gil-Jin Jang, Ho-Jin Choi, Minho Kim, Young-Won Kim, and Jaehun Choi. Recur- rent neural networks with missing information imputation for medical examination data prediction. In BigComp, pages 317-323, February 2017.","title":"Recurrent neural networks with missing information imputation for medical examination data prediction","authors":"Han-Gyu Kim, Gil-Jin Jang, Ho-Jin Choi, Minho Kim, Young-Won Kim, Jaehun Choi","year":"2017","venue":"BigComp","doi":null,"url":null,"internal_url":null},{"position":16,"raw":"[Kingma and Ba, 2015] Diederik Kingma and Jimmy Ba. Adam: A method for stochastic optimization. In ICLR, 2015.","title":"Adam: A method for stochastic optimization","authors":"Diederik Kingma, Jimmy Ba","year":"2015","venue":"ICLR","doi":null,"url":null,"internal_url":null},{"position":17,"raw":"[Kumar et al., 2006] Anand Kumar, Daniel Roberts, Ken- neth Wood, Bruce Light, Joseph Parrillo, Satendra Sharma, Robert Suppes, Daniel Feinstein, Sergio Zan- otti, Leo Taiberg, David Gurka, Aseem Kumar, and Mary Cheang. Duration of hypotension before initiation of ef- fective antimicrobial therapy is the critical determinant of survival in human septic shock. CCM, 34(6), June 2006.","title":"Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock","authors":"Anand Kumar, Daniel Roberts, Kenneth Wood, Bruce Light, Joseph Parrillo, Satendra Sharma, Robert Suppes, Daniel Feinstein, Sergio Zanotti, Leo Taiberg, David Gurka, Aseem Kumar, Mary Cheang","year":"2006","venue":"CCM","doi":null,"url":null,"internal_url":null},{"position":18,"raw":"[Lipton et al., 2016] Zachary Lipton, David Kale, and Ran- dall Wetzel. Directly modeling missing data in sequences with RNNs: Improved classification of clinical time series. JMLR, 56, 2016.","title":"Directly modeling missing data in sequences with RNNs: Improved classification of clinical time series","authors":"Zachary Lipton, David Kale, Randall Wetzel","year":"2016","venue":"JMLR","doi":null,"url":null,"internal_url":null},{"position":19,"raw":"[Maass, 1997] Wolfgang Maass. Networks of spiking neu- rons: The third generation of neural network models. Neu- ral Networks, 10(9), 1997.","title":"Networks of spiking neurons: The third generation of neural network models","authors":"Wolfgang Maass, Maass","year":"1997","venue":"Neural Networks","doi":null,"url":null,"internal_url":null},{"position":20,"raw":"[Neil et al., 2016] Daniel Neil, Michael Pfeiffer, and Shih- Chii Liu. Phased lstm: Accelerating recurrent network training for long or event-based sequences. In NIPS, 2016.","title":"Phased lstm: Accelerating recurrent network training for long or event-based sequences","authors":"Daniel Neil, Michael Pfeiffer, Shih-Chii Liu","year":"2016","venue":"NIPS","doi":null,"url":null,"internal_url":null},{"position":21,"raw":"[Parveen and Green, 2004] Shahla Parveen and Phil Green. Speech enhancement with missing data techniques using recurrent neural networks. In ICASSP, volume 1, pages I-733-736, May 2004.","title":"Speech enhancement with missing data techniques using recurrent neural networks","authors":"Green Parveen, Shahla Parveen, Phil Green","year":"2004","venue":"ICASSP","doi":null,"url":null,"internal_url":null},{"position":22,"raw":"[Polat et al., 2017] Gizem Polat, Rustem Ugan, Elif Cadirci, and Zekai Halici. Sepsis and septic shock: Current treat- ment strategies and new approaches. EJM, 49, 2017.","title":"Sepsis and septic shock: Current treatment strategies and new approaches","authors":"Gizem Polat, Rustem Ugan, Elif Cadirci, Zekai Halici","year":"2017","venue":"EJM","doi":null,"url":null,"internal_url":null},{"position":23,"raw":"[Reinhart et al., 2017] Konrad Reinhart, Ron Daniels, Niran- jan Kissoon, Flavia Machado, Raymond Schachter, and Simon Finfer. Recognizaing sepsis as a global health priority-a who resolution. NEJM, August 2017.","title":"Recognizaing sepsis as a global health priority-a who resolution","authors":"Konrad Reinhart, Ron Daniels, Niranjan Kissoon, Flavia Machado, Raymond Schachter, Simon Finfer","year":"2017","venue":"NEJM","doi":null,"url":null,"internal_url":null},{"position":24,"raw":"[Rubin and Little, 1987] Donald Rubin and Roderick Little. Statistical Analysis with Missing Data. John Wiley & Sons, Hoboken, NJ, 1987.","title":"Statistical Analysis with Missing Data","authors":"Donald Rubin, Roderick Little","year":"1987","venue":null,"doi":null,"url":null,"internal_url":null},{"position":25,"raw":"[Singer et al., 2016] Mervyn Singer, Clifford Deutschman, Christopher Seymour, Manu Shankar-Hari, Djillali Annane, Michael Bauer, Rinaldo Bellomo, Gordon Bernard, Jean-Daniel Chiche, Craig Coopersmith, Richard Hotchkiss, Mitchell Levy, John Marshall, Greg Martin, Steven Opal, Gordon Rubenfeld, Tom van der Poll, Jean- Louis Vincent, and Derek Angus. The third international consensus definitions for sepsis and septic shock (sepsis- 3). AMA, 315(8), February 2016.","title":"The third international consensus definitions for sepsis and septic shock (sepsis-3)","authors":"Mervyn Singer, Clifford Deutschman, Christopher Seymour, Manu Shankar-Hari, Djillali Annane, Michael Bauer, Rinaldo Bellomo, Gordon Bernard, Jean-Daniel Chiche, Craig Coopersmith, Richard Hotchkiss, Mitchell Levy, John Marshall, Greg Martin, Steven Opal, Gordon Rubenfeld, Tom Van Der Poll, Jean-Louis Vincent, Derek Angus","year":"2016","venue":"AMA","doi":null,"url":null,"internal_url":null},{"position":26,"raw":"[Tresp and Briegel, 1997] Volker Tresp and Thomas Briegel. A solution for missing data in recurrent neural networks with an application to blood glucose prediction. In NIPS, pages 971-977, 1997.","title":"A solution for missing data in recurrent neural networks with an application to blood glucose prediction","authors":"Briegel ; Tresp, Tresp Volker, Thomas Briegel","year":"1997","venue":"NIPS","doi":null,"url":null,"internal_url":null}],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}