What: Virtual lecture series on topics across machine learning in medicine, featuring extensive Q & A and panel discussions
Why: To reduce academia’s carbon footprint, accommodate the schedules of the world’s top scientists and maintain social distancing
Who: All are welcome!
Where: Zoom Webinar
Recordings of previous talks can be found here
Qingyu Zhao, PhD - Instructor, Department of Psychiatry and Behavioral Sciences, Stanford University
When: May 13th, 2022, 9:45-11:00 am
Title: Confounder-aware Deep Learning Models for Neuroimaging Applications
Abstract: The presence of confounding effects is one of the most critical challenges in applying deep learning techniques to medical applications. Confounders are extraneous variables that influence both input and output variables and thereby can easily distort the training and interpretation of deep learning models. How to remove confounding effects is a widely explored topic in traditional statistical research but is largely overlooked in the surge of deep learning applications as researchers put more attention on designing deeper and more powerful network architectures. In this talk, I will summarize our recent efforts in modeling confounding effects in deep learning models in the context of neuroimaging studies. I will first review a common practice in traditional (non-deep) machine learning studies that use general linear models to regress out confounding effects from deterministic features. Then I will discuss how to translate this residualization concept to the deep learning setting where the features are learned dynamically in an end-to-end fashion. Lastly, I will highlight the strength of these new approaches in deriving confounder-free latent representations of MRI data, correcting feature distributions with respect to multiple confounding variables, and generating unbiased interpretations of the model.
Bio: Dr. Zhao is an instructor in the Department of Psychiatry and Behavioral Sciences at Stanford University. He obtained his Ph.D. in computer science in 2017 from the University of North Carolina at Chapel Hill and was a postdoc and research scientist in the Stanford Psychiatry department. His research has been focusing on identifying biomedical phenotypes associated with neuropsychiatric disorders by statistical and machine-learning-based computational analysis of neuroimaging and neuropsychological data. Dr. Zhao is a recipient of the K99/R00 Pathway to Independence Award from the National Institute on Alcohol Abuse and Alcoholism.
Smita Krishnaswamy, PhD - Associate Professor of Genetics and Computer Science, Yale School of Medicine
When: May 27th, 2022, 9:45-11:00 am
Title: Deep Geometric and Topological Representations for Extracting Insights from Biomedical Data
Abstract: High-throughput, high-dimensional data has become ubiquitous in the biomedical sciences because of breakthroughs in measurement technologies. These large datasets, containing millions of observations of cells, molecules, brain voxels, and people, hold great potential for understanding the underlying state space of the data, as well as drivers of differentiation, disease, and progression. However, they pose new challenges in terms of noise, missing data, measurement artifacts, and the “curse of dimensionality.” In this talk, I will show how to leverage data geometry and topology, embedded within modern machine learning frameworks, to understand these types of complex scientific data. First, I will use data geometry to obtain representations that enable denoising, dimensionality reduction, and visualization. Next, I will show how to combine diffusion geometry with topology to extract multi-granular features from the data for predictive analysis. Then, I will move up from the local geometry of individual data points to the global geometry of data clouds and graphs, using graph signal processing to derive representations of these entities and optimal transport for distances between them. Finally, I will demonstrate how two neural networks use geometric inductive biases for generation and inference: GRASSY (geometric scattering synthesis network) for generating new molecules and molecular fold trajectories, and TrajectoryNet for performing dynamic optimal transport between time-course samples to understand the dynamics of cell populations. Throughout the talk, I will include examples of how these methods shed light on the inner workings of biomedical and cellular systems including cancer, immunology and neuroscientific systems.. I will finish by highlighting future directions of inquiry.
Bio: Smita Krishnaswamy is an Associate Professor in the department of Genetics and Computer Science at Yale, a core member of the Program in Applied Mathematics, Computational Biology and Interdisciplinary Neuroscience. She is also affiliated with the Yale Center for Biomedical Data Science, Yale Cancer Center, and Wu-Tsai Institute. Smita’s research focuses on developing deep representation learning methods that use mathematical concepts from manifold learning, data geometry, topology and signal processing. to denoise, impute, visualize and extract structure, patterns and relationships from big, high throughput, high dimensional biomedical data. Her methods have been applied to a variety of datasets from many systems in cancer biology, immunology, neuroscience, and structural biology.
Smita teaches three courses: Machine Learning for Biology (Fall), Deep Learning Theory and applications (spring), Advanced Topics in Machine Learning & Data Mining (Spring). She completed her postdoctoral training at Columbia University in the systems biology department where she focused on learning computational models of cellular signaling from single-cell mass cytometry data. She was trained as a computer scientist with a Ph.D. from the University of Michigan’s EECS department where her research focused on algorithms for automated synthesis and probabilistic verification of nanoscale logic circuits. Following her time in Michigan, Smita spent 2 years at IBM’s TJ Watson Research Center as a researcher in the systems division where she worked on automated bug finding and error correction in logic.