Causal modeling allows predicting a system’s behavior not only under observation but also under intervention. Computational causal discovery reverse-engineers causal models (networks) from observational data with limited or no interventions. In this work, I will present logic-based causal discovery, a new, versatile approach for learning causal networks from observations and interventions: based on standard causal assumptions, associative patterns in the data that constrain the search space of possible causal models are expressed as a logic formula.
Biomedical informaticians are inter-disciplinarians. This is no more evident than in the role of the CRIO, which requires engagement in and support to dozens of sub-fields across clinical and basic endeavors in the health sciences.
Machine learning is commonly described as a “field of study that gives computers the ability to learn without being explicitly programmed” (Simon, 2013). Despite this common claim, practitioners know that designing effective machine learning pipelines is often a tedious endeavor, and typically requires considerable experience with machine learning algorithms, expert knowledge of the problem domain, and brute force search to accomplish.
Saja Al-Alawneh abstract: Providing radiologists with feedback has been shown to improve their performance in mammography diagnosis. In 1992, the Mammography Quality Standards Act (MQSA) was enacted to improve the quality of mammography using audit and feedback procedures. However, no standard audit and feedback system for radiologists has been installed in the United States. Instead, auditing typically requires human effort to manually correlate radiology and pathology results.
Rathnam Abstract: Ubiquitin is arguable one of the most important molecules involved in post-translational modifications as it is present in all eukaryotic cells and plays a key role in mediating a wide assortment of biological processes, such as cell cycle regulation, endocytosis of cellular proteins, and transcriptional regulation.
Lee Abstract: The dysregulation of microRNAs (miRNAs) alters expression level of pro-oncogenic or tumor suppressive mRNAs in breast cancer, and in the long run, causes multiple biological abnormalities.
The long-term goal of the proposed work is to develop an effective informatics intervention that prevents harm to nursing home (NH) residents from potential drug-drug interactions (PDDIs) while avoiding known issues with interaction alerting such as alert fatigue. In this talk I will discuss progress toward that goal that examines the feasibility and potential clinical usefulness of actively monitoring patients exposed to psychotropic drugs.
Recent developments in scalable Bayesian inference have enabled fast learning of complex probabilistic models using massive data sets. We will discuss these developments in the context of topic models of discrete grouped data, focusing on text. We will review our recent collaborations on scalable model learning using stochastic variational inference, and discuss new applications to structured hierarchical topic models.