<div dir="ltr"><div class="gmail_quote"><div dir="ltr"><div>Team,</div><div><br></div><div>We will have a pleasure to watch Chirag give his thesis proposal talk in less than an hour. When the time comes, I recommend that you grab your dinners, sit comfortably in front of your computer screens, and enjoy the talk.</div><div><br></div><div>See below for the zoom link, etc.</div><div><br></div><div>Cheers</div><div>Artur</div><div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">---------- Forwarded message ---------<br>From: <strong class="gmail_sendername" dir="auto">Stacey Young</strong> <span dir="auto"><<a href="mailto:StaceyYoung@cmu.edu" target="_blank">StaceyYoung@cmu.edu</a>></span><br>Date: Mon, Dec 6, 2021 at 4:45 PM<br>Subject: LTI PhD Thesis Proposal: Chirag Nagpal - Wednesday, December 15, 2021 (6pm)<br>To: <<a href="mailto:lti-students@cs.cmu.edu" target="_blank">lti-students@cs.cmu.edu</a>>, <<a href="mailto:lti-staff@cs.cmu.edu" target="_blank">lti-staff@cs.cmu.edu</a>>, LTI-faculty-all <<a href="mailto:lti-faculty-all@cs.cmu.edu" target="_blank">lti-faculty-all@cs.cmu.edu</a>><br></div><br><br><div dir="ltr">The LTI is proud to announce the following PhD Thesis Proposal:<div><br></div><div>Leveraging Heterogeneity in Time-to-Event Predictions</div><div><br></div><div>Chirag Nagpal</div><div><br></div><div>Wednesday, December 15, 2021</div><div>6:00pm via <a href="https://cmu.zoom.us/j/93544413242?pwd=SkJIVW5vMkdoMVdHUnpWSzVtZ0t1Zz09" target="_blank">Zoom</a><br><b>ID</b>: 93544413242<br><b>Password</b>: 024906</div><div><br></div><div><b>Committee:</b></div><div><br></div><div>Artur Dubrawski, Robotics Institute, (Chair)</div><div>Louis-Philippe Morency<br>Bhiksha Ramakrishnan</div><div>Russell Greiner, (University of Alberta)<br>Katherine Heller, (Duke University and Google Research)</div><div> <br></div><div><b>Abstract:</b></div><div><br></div><div>Time-to-Event Regression, often referred to as Survival Analysis or Censored Regression involves learning of statistical estimators of the survival distribution of an individual given their covariates. As opposed to standard regression, survival analysis is challenging as it involves accounting for outcomes censored due to loss of follow up. This circumstance is common in, e.g., bio-statistics, predictive maintenance, and econometrics. With the recent advances in machine learning methodology, especially deep learning, it is now possible to exploit expressive representations to help model survival outcomes. My thesis contributes to this new body of work by demonstrating that problems in survival analysis often manifest inherent heterogeneity which can be effectively discovered, characterized, and modeled to learn better estimators of survival. <br>Heterogeneity may arise in a multitude of settings in the context of survival analysis. Some examples include heterogeneity in the form of input features or covariates (for instance, static vs. streaming, time-varying data), or multiple outcomes of simultaneous interest (more commonly referred to as competing risks). Other sources of heterogeneity involve latent subgroups that manifest different base survival rates or diverse responses to an intervention or treatment. Additional sources of heterogeneity involve shifts in the covariate or outcome distributions at the time of model deployment, exacerbating risk of mis-estimation. Furthermore, heterogeneity may manifest in the outcomes of protected demographics, especially if group membership information is not available or misreported. <br>In this thesis, I aim to demonstrate that carefully modelling the inherent structure of heterogeneity can boost predictive power of survival analysis models while improving their specificity and precision of estimated survival at an individual level. An overarching methodological framework of this thesis is the application of graphical models to impose inherent structure in time-to-event problems that explicitly model heterogeneity, while employing advances in deep learning to learn powerful representations of data that help leverage various aspects of heterogeneity.<br></div><div><br></div><div>A copy of the thesis proposal can be found <a href="https://drive.google.com/file/d/1-uQ0w5x-zoY6mSXrU3aahGW2VWEAoLdJ/view" target="_blank">here</a>.</div><div><br></div><div><br clear="all"><div><br></div>-- <br><div dir="ltr" data-smartmail="gmail_signature"><div dir="ltr"><pre cols="72" style="white-space:pre-wrap;color:rgb(34,34,34)">Stacey L. Young
PhD Academic Program Manager
LTI PhD Program, LTI Dual PhD/Portugal Program
LTI Minor/Concentration & Admin Support Robert Frederking
Language Technologies Institute
Pronouns she/her
Black Lives Matter
6415 Gates Hillman Complex
5000 Forbes Avenue
Pittsburgh, PA 15213</pre><pre cols="72" style="white-space:pre-wrap;color:rgb(34,34,34)">T: (412)268-2623
F: (412)268-6298</pre><pre cols="72" style="white-space:pre-wrap;color:rgb(34,34,34)"><b>Office Hours:
</b><b>Remote: </b>M-TH, 10am-5:30pm<b>
In Person:</b> F, 10am-5:30pm</pre></div></div></div></div>
</div><br clear="all"><div><br></div>-- <br><div dir="ltr" data-smartmail="gmail_signature"><div dir="ltr"><div><div dir="ltr"><div><div dir="ltr"><span style="color:rgb(0,0,0)"><span style="background-color:rgb(255,255,255)"><span><b>Chirag Nagpal<br></b>
PhD Student, Auton Lab<br>
School of Computer Science<br>
Carnegie Mellon University<br></span></span></span></div><div><span style="color:rgb(0,0,0)"><span style="background-color:rgb(255,255,255)"><span><a href="http://cs.cmu.edu/~chiragn" target="_blank">cs.cmu.edu/~chiragn</a><br></span></span></span></div></div></div></div></div></div></div></div>
</div></div>