=Paper=
{{Paper
|id=Vol-2551/keynote-01
|storemode=property
|title=None
|pdfUrl=https://ceur-ws.org/Vol-2551/keynote-01.pdf
|volume=Vol-2551
|dblpUrl=https://dblp.org/rec/conf/wsdm/Lipton20
}}
==None==
<pdf width="1500px">https://ceur-ws.org/Vol-2551/keynote-01.pdf</pdf>
<pre>
       Machine Learning for Healthcare: Beyond i.i.d. Prediction
                                                         Zachary C. Lipton
                                                      Carnegie Mellon University
                                                         Pittsburgh, PA, USA
                                                          zlipton@cmu.edu

ABSTRACT                                                               quantities, uncertainty quantification is often essential, and individ-
Following breakthroughs in computer vision and natural language        uals are characterized by data from multiple modalities. In this talk,
processing, widespread excitement and financial support have buoyed    I will discuss recent breakthroughs in deep learning for healthcare,
a rapidly-growing field of machine learning for healthcare. And yet,   as well as my own group’s work pioneering RNNs for multivariate
while most of machine learnings most impressive results concern        clinical time series data and then focus on our more recent efforts to
point estimates under strict i.i.d. assumptions, medical decision-     address aspects of decision-making that are fundamentally missing
making often requires something more. Conditions shift (due to         in the standard machine learning setup.
seasonality, changing prevalence of illnesses, and availability of
tests), the quantities of interest are often counterfactual (causal)
                                                                       Copyright © 2020 for this paper by its authors. Use permitted under Creative Commons
                                                                       License Attribution 4.0 International (CC BY 4.0).

</pre>