<div dir="ltr">fysa, Rohini's MS defense is on Tuesday next week.<div><br></div><div>Cheers,</div><div>Artur<br><br><div class="gmail_quote gmail_quote_container"><div dir="ltr" class="gmail_attr">---------- Forwarded message ---------<br>From: <strong class="gmail_sendername" dir="auto">Tracy Farbacher</strong> <span dir="auto"><<a href="mailto:tt1b@andrew.cmu.edu">tt1b@andrew.cmu.edu</a>></span><br>Date: Tue, Apr 15, 2025 at 7:18 PM<br>Subject: 5th Year MS thesis Presentation Rohini Banerjee<br></div><br><div dir="ltr"><div><b>Rohini Banerjee<br>Tuesday, April 22, 2025<br>3:00 PM<br>Newell Simon Hall (NSH) 3001<br>Uncertainty-Aware AI for Clinical Decision Support</b></div><div><br><div><b>Abstract:</b></div><div>Building
interpretable-by-design AI models that intuitively communicate model
uncertainty is vital to engendering physician and patient trust. We
develop uncertainty-guided deep learning systems for two pertinent
healthcare settings. Efficient intravascular access in trauma and
critical care is a high-stakes intervention affording minimal tolerance
for error. Autonomous needle insertion systems can be useful in austere
environments due to the lack of skilled medical personnel. However,
inaccuracies in vessel segmentation modeling can result in vessel damage
and hemorrhage. The risk can be mitigated via predictive uncertainty
estimation to assess model reliability. Thus, we introduce MSU-Net, a
novel multistage approach to semantic vessel segmentation in ultrasound
images that combines the predictive power of Monte Carlo networks and
deep ensembles. We demonstrate significant improvements, 27.7% over the
state-of-the-art, while enhancing model reliability through a 20.9%
stronger discrimination in epistemic uncertainty between correct and
incorrect predictions.</div><div><br>Next, we investigate the robustness
of predictive modeling in quantifying the severity of rash
manifestations associated with Cutaneous Dermatomyositis (CDM), a rare
and currently incurable autoimmune disorder. Given the importance of
telemedicine for remote disease monitoring and timely intervention, we
address challenges of data scarcity and patient diversity by integrating
a novel BERT-style self-supervised learning (SSL) framework to
image-based models. Pretrained via masked image modeling on
demographically diverse images, our model achieves over a 40%
improvement in fine-tuning performance on high-resolution in-clinic hand
images from a limited cohort of 23 CDM patients. We achieve 83%
accuracy on a held-out patient set, surpassing the clinical benchmark of
70–75% accuracy. To our knowledge, this is the first work to integrate
uncertainty estimation into such architectures, enabling robustness
under distributional shift in skin tone unseen during fine-tuning.</div><div><br>Our
contributions lay the groundwork for developing accurate, statistically
rigorous, clinically actionable deep learning models. Future work aims
to improve the interpretability of models for equitable clinical
decision support.</div><div><b><br></b></div><div><b>Thesis Document:</b></div><div><a href="https://drive.google.com/file/d/1TnC6IcyDmTD6IzBpaSKsIsQX1f8Bkl0K/view?usp=sharing" target="_blank">https://drive.google.com/file/d/1TnC6IcyDmTD6IzBpaSKsIsQX1f8Bkl0K/view?usp=sharing</a></div><div><br></div><div><b>Thesis Committee:</b></div><div>Artur W. Dubrawski (Chair) </div><div>László A. Jeni<br><br></div><br></div></div>
</div></div></div>