Schedule

Session 1: Introduction (Chair: Kyunghyun Cho)
9.00-9.20 Overview of the Center for Data Science by Claudio Silva (CDS Director)
9.20-9.40 Overview of the Brain Initiative at NYU by Bijan Pesaran (Center for Neural Science)
9.40-10.10 Overview of the Data Science Activities at Langone by Dan Sodickson (Center for Biomedical Imaging) and David Fenyo (Institute for Systems Genetics)

10.10-10.30 Coffee Break

Session 2: Research Problems in Medicine and Health (Chair: Dan Sodickson)
10.30-11.00 Talk 1 Florian Knoll (Radiology)
11.00-11.30 Talk 2 Narges Razavian (Population Health)
11.30-12.00 Talk 3 Kelly Ruggles (Applied Bioinformatics Laboratories)

12.00-13.25 Poster session + Lunch

Session 3: Lightning talks (Chair: David Fenyo)
13.25 – 14.10: Nine 5-min spotlight talks

14.10-14.40 Coffee Break

Session 4: Data Science Research (Chair: Rumi Chunara)
14.40-15.10 Talk 1 Guido Gerig (Tandon Computer Science Engineering)
15.40-16.10 Talk 2 Uri Shalit (Courant Computer Science)
15.40-16.10 Talk 3 Kyunghyun Cho

Session 5: Brainstorming with Panelists
16.10-17.30 Panel Discussion
Moderator: Narges Razavian
Panelists:
Rumi Chunara (Global Public Health)
Aristotelis Tsirigos (Department of Pathology)
Yindalon Aphinyanaphongs (Center for Predictive Analytics)
Itai Yanai (Institute for Computational Medicine)

Talks

Florian Knoll

Title: Medical Image Reconstruction: Why should a Data Scientist care? (Opportunities and Challenges for Machine Learning)

Abstract: This talk will provide an introduction to the use of machine learning and convolutional neural networks (CNNs) in the field of MR image reconstruction. We will use the example of reconstruction from undersampled data from accelerated acquisitions throughout the talk and will base our formulation on iterative reconstruction methods as used in compressed sensing (CS). We will formulate a CNN based reconstruction that can be seen as a generalization of CS, and explain how we can learn an entire image reconstruction procedure. Using selected examples, we will discuss both advantages and challenges, covering topics like reconstruction time, design of the training procedure, error metrics, training efficiency and validation of image quality.

Bio: Dr. Knoll is an Assistant Professor at the Center for Biomedical Imaging at the Department of Radiology of the NYU School of Medicine. His main research interest lies in the development of reconstruction methods for medical imaging. Using developments from the fields of inverse problems, variational optimization and machine learning, he is working on image reconstruction for accelerated Magnetic Resonance (MR) imaging, the synergistic combination of multimodality imaging data at the image reconstruction stage, quantitative MR imaging and low dose Computed Tomography (CT).

Narges Razavian

Title: Machine Learning for Population Health: Data Landscape and Research Opportunities

Abstract: Electronic health records have been accumulating in medical centers for a few years now and growing every year, opening doors to data-driven hypothesis generation and novel machine learning applications. In this talk, we will review data modalities and quantities currently available at New York University Langone Medical Center, and briefly introduce a subset of research projects and collaborations focused on machine learning and disease modeling. These research directions include prediction of future risk of 133 diseases from temporal trends on the lab results using temporal convolution, automatic detection of chronic heart failure for the hospitalized patients, disease diagnosis and subtype detection using MRI data, predicting physicians’ behavior and errors, and prediction of childhood obesity. We close by reviewing avenues for further collaborations between medical and computational researchers.

Bio: Narges Razavian, PhD, is an assistant (research) professor in the predictive analytics group in the Department of Population Health and the Center for Healthcare Innovation and Delivery Science. Her research focuses on bridging the gap between machine learning and healthcare, in particular, using most advanced computational models to help with clinical research, including, but not limited to, biomarker discovery, early detection, care quality control, patient similarity, and prediction of future outcomes. From the machine learning point of view, she is excited to work with diverse patient information and high dimensional data (electronic health records, images, text, social, genomics), especially in forms of time series.

Kelly Ruggles

Title: Identifying Therapeutic Targets in Breast Cancer using Proteogenomics

Abstract: Cancer has been well established as a disease of the genome, with a subset of somatic mutations frequently acting as drivers of tumor progression, and thereby influencing diagnosis, prognosis and treatment. The integration of cancer genomics with mass spectrometry-based proteomics can be used to supplement genomic information, determining the effect of genomic aberrations at the protein level, guiding biomarker identification and predicting effective drug combinations for treatment. Our group focuses on the development and application of computational methods applied to quantitative proteomic, transcriptomic, and genomics data to get at these important questions. During the seminar, I will discuss a subset of these methods, which we have used to identify promising treatment strategies and potential biomarkers in breast cancer. Additionally, I will discuss how we are expanding these efforts to study diverse tumor and sample types using a similar approach.

Bio: Kelly Ruggles is an Assistant Professor in the Department of Medicine and a member of the Applied Bioinformatics Laboratories at NYU Langone Medical Center. Her research focuses on the integration of diverse data modalities and development of complementary statistical analysis methods to uncover important biological insights from highly complex datasets. Although her research topics are broad, including cancer proteomics, next generation sequencing, and microbiome analysis, her primary interests lie in developing data-type agnostic analysis methods to bridge the gap between biology and data science.

Guido Gerig

Title: Medical Image Analysis: Extracting Information from Image Data

Abtract: Medical image analysis is a multidisciplinary field at the intersection of computer and data science, engineering, mathematics, statistics and medicine. With the main goal of extracting clinically relevant information from medical image data, we develop mathematical and computational methods to solve problems related to the use of imaging for biomedical research and clinical applications. This talk demonstrates how progress in fundamental image analysis methodologies enables new discoveries in medicine and helps to translate novel techniques to clinical routine. In particular, methodologies for studies of dynamic processes via longitudinal imaging will be discussed. We will present typical examples of interdisciplinary imaging research projects, covering clinical research areas as diverse as autism, traumatic brain injury, ophthalmology and diagnostic radiology.

Bio: Guido Gerig joined the Computer Science and Engineering (CSE) Department at NYU Tandon School of Engineering in 2015 as Institute Professor and a member of the Visualization, Imaging and Data Analysis (VIDA) Center. In 2016, he also took on the responsibility as department chair of CSE. He was previously USTAR professor at the School of Computing and Associate Director of the Scientific Computing and Imaging Institute at the University of Utah. He is a fellow of the Medical Image Computing and Computer-Assisted Intervention (MICCAI) Society and of the American Institute for Medical and Biological Engineering (AIMBE), and he serves on the executive committee of the Elsevier Journal Medical Image Analysis. Guido Gerig’s research is centered about development of new methods and techniques for medical image analysis and computer vision, with a strong emphasis on collaborative research with clinical groups. Major research topics include analysis of single and multi-timepoint volumetric imaging data, longitudinal modeling of shape and image data, joint analysis of geometry and appearance from 4D image data, and segmentation/registration/modeling methods for image data presenting novel appearing or disappearing structures as presented by large pathologies. Driving clinical applications include neurodevelopmental aspects in autism, infants at risk for mental illness, neurodegeneration in Huntington’s disease and pathologies such as presented in severe TBI and tumor. Guido Gerig is author/co-author of over 300 articles in international peer-reviewed journals and conferences. Image analysis tools and methods developed by his group are distributed to the public as open-source tools.

Uri Shalit

Title: Deep learning approaches for individual level causal inference

Abstract: Machine learning has had enormous success in the last few years addressing prediction and detection problems, such as automatically detecting skin cancer from images of skin lesions. However, many tasks in medicine require not just detection or prediction, but action. Examples include prescribing medication or deciding to move a patient to an ICU. I will discuss how learning to act is different from learning to detect or predict, and how is learning to act related to causal inference. I will then present several deep-learning based methods for causal inference developed in our group and discuss their applications.

Bio: Uri Shalit is a postdoctoral researcher in the Courant Institute of Mathematical Sciences, New York University, working at David Sontag’s Clinical Machine Learning Lab. His research is focused on creating new methods for finding causal relationships in large-scale high-dimensional observational studies, motivated by applications in healthcare and clinical medicine. Uri completed his PhD studies at the School of Computer Science & Engineering at The Hebrew University of Jerusalem, under the guidance of Prof. Gal Chechik and Prof. Daphna Weinshall. From 2011 to 2014 Uri was a recipient of Google’s European Fellowship in Machine Learning. Uri will be joining the faculty of the Technion Israel Institute of Technology in the Fall of 2017.

Kyunghyun Cho

Title: The Workshop

Abstract: I will describe why and how this workshop has been organized, and what the goals of this workshop are. I will then discuss the future directions the organizers hope this workshop would lead to.

Bio: Kyunghyun Cho is an assistant professor of computer science and data science at New York University. He was a postdoctoral fellow at University of Montreal until summer 2015 under the supervision of Prof. Yoshua Bengio, and received PhD and MSc degrees from Aalto University early 2014 under the supervision of Prof. Juha Karhunen, Dr. Tapani Raiko and Dr. Alexander Ilin. He tries his best to find a balance among machine learning, natural language processing and life, but almost always fails to do so.

Panelists

Claudio Silva

Claudio T. Silva is a professor of computer science and engineering and data science at New York University. His research interests include visualization, visual analytics, reproducibility and provenance, geometric computing, data science/big data, sports analytics, urban computing and computer graphics. He has held positions in academia and industry, including at AT&T, IBM, Lawrence Livermore National Labs, Sandia National Labs, and the University of Utah. Claudio has advised 15 PhD and 8 MS students, and mentored 6 post-doctoral associates. He has published over 220 journal and conference papers, is an inventor of 12 US patents, and authored 12 papers that have received “Best Paper Awards” (including honorable mentions). He has over 10,000 citations according to Google Scholar. He is an IEEE Fellow and was the recipient of the 2014 IEEE Visualization Technical Achievement Award “in recognition of seminal advances in geometric computing for visualization and for contributions to the development of the VisTrails data exploration system.” Claudio’s research has been funded by NSF, DOE, NIH, NASA, DOD, AT&T, IBM, ExxonMobil, McGraw-Hill Education, MLBAM, Moore Foundation, Sloan Foundation, LLNL, Sandia, Los Alamos, State of Utah, University of Utah, Center for Urban Science and Progress, and New York University.

Rumi Chunara

Dr. Chunara is an Assistant Professor at NYU, jointly appointed at the Tandon School of Engineering (in Computer Science) and the College of Global Institute of Public Health. Her research interests are in population-level disease surveillance. She specializes in building and harnessing data from participatory tools, such as point of care diagnostics, mobile phones and other Internet-enabled sensors and media. Simultaneously, Dr. Chunara develops statistical and machine learning methodology for using these observational data sources in epidemiological models. Dr. Chunara completed her PhD at the Harvard-MIT Division of Health, Sciences and Technology, Master’s degree at MIT in Electrical Engineering and Computer Science and received her Bachelor’s degree in Electrical Engineering from Caltech with honors. She is a recipient of the MIT Presidential Fellowship and a Caltech Merit Scholarship, the NYC Media Lab – Bloomberg Data for Good Exchange Paper Award and was selected as an MIT Technology Review Top 35 Innovator under 35 in 2014.

Yindalon Aphinyanaphongs

I am an assistant professor in the Center for Health Informatics and Bioinformatics. My research involves understanding how to identify social media content automatically using machine learning, theoretical work on machine learning and text classification, and health information flows within social networks. We aim to build on the insights from social media work in political science toward applications in healthcare.

Itai Yanai

Dr. Itai Yanai joined the faculty at New York University’s School of Medicine in May 2016 as a professor in the Department of Biochemistry and Molecular Pharmacology. He serves as the inaugural director of the Institute for Computational Medicine (ICM), whose goal is to harness computational approaches for fundamental and medically-relevant discoveries. Through the development of novel tools, the nurturing of young investigators, and translational applications, ICM aims to create a culture that promotes scientific advancements.

Dr. Yanai’s research focuses on the interface of gene expression, development, and evolution. Using his training as an experimental embryologist, a molecular biologist, and a computational biologist, his interest is exploring how developmental pathways evolve at the molecular level. Members of his lab carry out intricate embryological experiments at the level of individual cells and apply computational approaches to explore the resulting data. As a model system, they use the best understood animal, the nematode C. elegans. His lab developed the popular CEL-Seq method for single-cell RNA-Seq and they have used it to study stages, germ-layers, and body-plans in animal embryos. More recently, his lab is applying single-cell RNA-Seq to the study of tumorigenesis and bacterial infection.

Dr. Yanai received his undergraduate degrees in Computer Engineering and the Philosophy of Science and his PhD in Bioinformatics from Boston University in 1997 and 2002, respectively. He completed a postdoctoral fellowship in Molecular Genetics in 2004 at the Weizmann Institute of Science in Israel and a postdoctoral fellowship in Developmental Genetics at Harvard University in 2008. He served as adjunct Assistant Professor of Bioinformatics at Boston University from 2004-2008. At the Technion–Israel Institute of Technology, he served as an Assistant Professor in the Department of Biology from 2008-2013 and Associate Professor from 2014-2016. He was a Radcliffe Fellow, Radcliffe Institute for Advanced Study, Harvard University, and a visiting professor, Broad Institute of Harvard and MIT, from 2014-2015.

In addition to his research goals, Dr. Yanai firmly believes that the communication of knowledge is a major component of science and is involved in mentoring students, giving presentations, participating in outreach programs and in the dissemination of science to a popular audience. Towards this end, Dr. Yanai has also recently co-authored a popular science book, entitled “The Society of Genes”, along with Dr. Martin Lercher from Heinrich-Heine University in Düsseldorf.

Aristotelis Tsirigos

Dr. Tsirigos is an Associate Professor of Pathology and the Director of the Applied Bioinformatics Center at the NYU School of Medicine. He received his Ph.D. in Computer Science from Courant Institute of New York University in January 2006, and his B.S. from the National Technical University of Athens, Greece in 1998. Dr. Tsirigos’ research focus is primarily in Computational Genomics and involves the design of statistical/computational methods and pipelines to model biological systems and generate novel biological hypotheses.