R25 VOICE Yin Aphinyanaphongs

Section 1 podcast.

Papers discussed in this Section 2 podcast:
Domingos, Pedro. 2012. “A Few Useful Things to Know about Machine Learning.” Communications of the ACM 55 (10):78.Luo W, Phung D, Tran T, Gupta S, Rana S, Karmakar C, Shilton A, Yearwood J, Dimitrova N, Ho TB, Venkatesh S, Berk M “Guidelines for Developing and Reporting Machine Learning Predictive Models in Biomedical Research: A Multidisciplinary View” J Med Internet Res 2016;18(12):e323 DOI: 10.2196/jmir.5870Podcast Contents:
GeneralizationOverfittingFeature EngineeringImproving model performanceMore dataBetter algorithmsEnsemblingReview of checklists for writing machine learning papers.Student questionsKnowledge vs DataJMIR reputationInformatics journals and computer science proceedingsSample size for good classifier performance.

Papers discussed in this Section 3 podcast:
Liao, Fangzhou; Liang, Ming; Li, Zhe; Hu, Xiaolin; and Song, Sen. Evaluate the Malignancy of Pulmonary Nodules Using the 3D Deep Leaky Noisy-or Network. eprint arXiv:1711.08324, 2017Pollard, T. J., & Johnson, A. E. W. The MIMIC-III Clinical Database. http://dx.doi.org/10.13026/C2XW26 (2016)Pranav Rajpurkar, Jeremy Irvin, Aarti Bagul, Daisy Ding, Tony Duan, Hershel Mehta, Brandon Yang, Kaylie Zhu, Dillon Laird, Robyn L. Ball, Curtis Langlotz, Katie Shpanskaya, Matthew P. Lungren, and Andrew Ng. MURA Dataset: Towards Radiologist-Level Abnormality Detection in Musculoskeletal Radiographs. arXiv:1712.06957, 2017X. Wang, Y. Peng, L. Lu, Z. Lu, M. Bagheri, R. M. Summers. ChestX-ray8: Hospital-scale Chest X-ray Database and Benchmarks on Weakly-Supervised Classification and Localization of Common Thorax Diseases. IEEE CVPR (spotlight);  arXiv:1705.02315, 2017Podcast Contents:
Why Datasets are important?Kinds of Datasets?What's a gold standard?Best practices in dataset descriptions.Sample distributionMeta-dataPatientsRadiologistsPACS Systems Used for AnnotationImagesStrategies for Labeling DataNatural Language ProcessingAmazon Mechanical TurkNatural Language Processing Validation Sets 

Papers discussed in this Section 4 podcast:
Anand Avati, Kenneth Jung, Stephanie Harman, Lance Downing, Andrew Ng, Nigam H. Shah. Improving Palliative Care with Deep Learning. arXiv:1711.06402Frizzell JD, Liang L, Schulte PJ, Yancy CW, Heidenreich PA, Hernandez AF, Bhatt DL, Fonarow GC, Laskey WK. Prediction of 30-Day All-Cause Readmissions in Patients Hospitalized for Heart Failure Comparison of Machine Learning and Other Statistical Approaches. JAMA Cardiol. 2017;2(2):204–209. doi:10.1001/jamacardio.2016.3956Joseph Futoma, Sanjay Hariharan, Mark Sendak, Nathan Brajer, Meredith Clement, Armando Bedoya, Cara O'Brien, Katherine Heller. An Improved Multi-Output Gaussian Process RNN with Real-Time Validation for Early Sepsis Detection. arXiv:1708.05894Riccardo Miotto, Li Li, Brian A. Kidd & Joel T. Dudley. Deep Patient: An Unsupervised Representation to Predict the Future of Patients from the Electronic Health Records. Scientific Reports 6, Article number: 26094 (2016) doi:10.1038/srep26094Podcast Contents:
Why These Papers?Predict 30 day all cause readmissionHow I was surprised.Appreciation for data inputs.Improving the classificationBetter representation through deep learning.Consider time rather than a snapshot of a given admission.Consider severity of the diseases.Consider medication dosages as a proxy for disease severity.Palliative CareObservation WindowsArea under the Precision Recall  Curve.The target is a proxy.Model explanation.Deep patientBuilding good features.Dealing with noisy data.Sparsity in the number of notes per patient.Sparsity in the number of patients with a feature.Topic Modeling.ICD-9 Granularity.ToolsOpen Biomedical AnnotatorEarly SepsisUndefined time zero.Dealing with time series.irregularly spaced recording.Informed missingness.Case control matching.Matched lookback.Realtime validation.Student Questions

Papers discussed in this Section 5 Podcast:
Andre Esteva, Brett Kuprel, Roberto A. Novoa, Justin Ko, Susan M. Swetter, Helen M. Blau & Sebastian Thrun. Dermatologist-level classification of skin cancer with deep neural networks. Nature 542, 115–118 (02 February 2017) doi:10.1038/nature21056Gulshan V, Peng L, Coram M, Stumpe MC, Wu D, Narayanaswamy A, Venugopalan S, Widner K, Madams T, Cuadros J, Kim R, Raman R, Nelson PC, Mega JL, Webster DR. Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs. JAMA. 2016;316(22):2402–2410. doi:10.1001/jama.2016.17216Pranav Rajpurkar, Jeremy Irvin, Kaylie Zhu, Brandon Yang, Hershel Mehta, Tony Duan, Daisy Ding, Aarti Bagul, Curtis Langlotz, Katie Shpanskaya, Matthew P. Lungren, Andrew Y. Ng. CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning. arXiv:1711.05225Podcast Contents
Why These Papers?Dermatology PaperConceptsInception v3Pretrainingt-SNEComparison to humansCombining clinical data with imagingCheXnetConceptsDensenetPretrainingHorizontal flippingClass Activation MappingsImplications of downscalingRetinopathy PaperHuman ComparisonDifferent CamerasConceptsPretrainingMultitask -single network, multiple outputsEarly stopping criteriaEnsembleLearning CurvesWhat is the Model Learning?

Papers discussed in this Section 6 Podcast:
H. Salehinejad, J. Barfett, S. Valaee, E. Colak, A. Mnatzakanian, and T. Dowdell. Interpretation of mammogram and chest radiograph reports using deep neural networks-preliminary results. arXiv preprint arXiv:1708.09254, 2017.Hassanpour S, Langlotz CP. Predicting High Imaging Utilization Based on Initial Radiology Reports: A Feasibility Study of Machine Learning. Acad Radiol 2016; 23 (01) 84-89.Pons E., Braun L.M.M., Hunink M.G.M. et al. (2016) Natural language processing in radiology: a systematic review. Radiology, 279, 329–343.Trivedi, H., Mesterhazy, J., Laguna, B. et al. Automatic Determination of the Need for Intravenous Contrast in Musculoskeletal MRI Examinations Using IBM Watson’s Natural Language Processing Algorithm. J Digit Imaging (2017). https://doi.org/10.1007/s10278-017-0021-3Podcast Contents
Why These PapersNLP ReviewDefining NLPNLP Pipeline in Figure 1RadlexEvaluation Measures - F1TypesDiagnostic SurveillanceCohort BuildingQuery based case retrievalQuality Assessment in radiologic practiceCommunication of critical resultsClinical Support ServicesResources in Table 2Operational BarriersFuture Research NeedsIV ContrastWhy Chosen?NotesProcessing Time DiscussionError analysisCloud ServicePassive Workflow integration.Predicting High Imaging UtilizationWhy Chosen?NotesSVM usage.Document-Feature MatrixOverfitInterpretation of MammogramsWhy Chosen?NotesBi-directional CNNPassive Workflow IntegrationPreprocessingWhy Deep LearningQuestions