Presenter: Michael Becich, MD, PhD/Jonathan Silverstein, MD/Shyam Visweswaran, MD, PhD

DBMI’s Research Informatics Office (RIO, see

Presenter: Franz Franchetti, PhD

In this talk I will first discuss the state of computing technology and the fast pace of change due to Moore's law, and set this into context of the time scale of other computing developments like algorithms and computer languages. I will show the impact on solvable computing problems at various computing scales.

Presenter: Xinghua Lu, MD, PhD

Cancer is mainly caused by somatic genome alterations that perturb cellular signaling pathways, and it is anticipated that precisely targeting patient-specific genomic alterations of individual tumors (precision oncology) will lead to more effective therapies.

Presenter: Eric Nyberg, PhD

This talk presents recent results on automatic question answering for biomedical and consumer health questions, drawn from CMU's participation in several recent open evaluations (including BioASQ, LiveQA, and MediQA). Despite promising achievements in recent evaluation campaigns, significant challenges must be solved before QA can be practically useful for experts and consumers.

Presenter: Haohan Wang, MS

The talk will start with the introduction of the most basic building blocks of deep learning models (especially convolutional neural networks) to build some statistical intuition of what deep learning is supposed to be capable of, and will show some evidence supporting that what is behind the promised human-level understanding of data is partially the model's tendency to capture the high-frequent superficial signals.

Presenter: Danielle Brown and Victoria Khersonsky
Presenter: Greg Cooper, MD, PhD

This talk describes an instance-specific causal Bayesian network (CBN) learning method that searches the space of CBNs to build a causal model that is specific to an instance (e.g., a patient). The search is guided by attributes of the given instance (e.g., patient symptoms, signs, lab results, and genotype). We describe the results of applying the method to molecular cancer data to estimate the gene alterations (e.g., gene mutations) that are driving the cancerous behavior of individual tumors, which are the instances in this application.