Keynote Speakers
SVP of AI and Machine Learning in Optum Labs (United Health Group), and Professor in the Departments
of Computer Science, Computational Medicine, Anesthesiology, and Human Genetics at UCLA
TITLE: Whole-genome Methylation Patterns as a Biomarker for EHR Imputation
ABSTRACT: Diagnosis and prediction of health outcomes using machine learning has shown major advances over the last few years. One of the challenges in applying machine learning models to electronic health records is that the data is sparse and often noisy. To address sparsity, imputation methods are typically applied. These approaches use the correlation structure between different medical features to impute the missing information. Arguably, genomic information can add a useful set of features that would improve the imputation accuracy of the medical records. It has been suggested that genetic information, and particularly polygenic risk, is often predictive, particularly for highly heritable conditions. In this talk, I will describe an evaluation of methylation patterns (modifications to the DNA) as a predictive tool that can be used for EHR imputation. We show that methylation provides a better imputation performance when compared to genetic or EHR data. Our approach uses a new tensor deconvolution of bulk DNA methylation to obtain cell-type-specific methylation that is in turn used for imputation.
BIO: Eran Halperin is the SVP of AI and Machine Learning in OptumLabs (United HealthGroup), and a professor in the departments of Computer Science, Computational Medicine, Anesthesiology, and Human Genetics at UCLA. Prior to his current position, he held research and postdoctoral positions at the University of California, Berkeley, the International Computer Science Institute in Berkeley, Princeton University, and Tel-Aviv University. Dr. Halperin’s lab developed methods for a variety of health-related domains, including different genomic domains (genetics, methylation, microbiome, single-cell RNA), and medical applications (medical imaging, physiological waveforms, and electronic medical records). He published more than 150 peer-reviewed publications, and he received various honors for academic achievements, including the Rothschild Fellowship, the Technion-Juludan prize for technological contribution to medicine, the Krill prize, and he was elected as an International Society of Computational Biology (ISCB) fellow.
of Computer Science, Computational Medicine, Anesthesiology, and Human Genetics at UCLA
TITLE: Whole-genome Methylation Patterns as a Biomarker for EHR Imputation
ABSTRACT: Diagnosis and prediction of health outcomes using machine learning has shown major advances over the last few years. One of the challenges in applying machine learning models to electronic health records is that the data is sparse and often noisy. To address sparsity, imputation methods are typically applied. These approaches use the correlation structure between different medical features to impute the missing information. Arguably, genomic information can add a useful set of features that would improve the imputation accuracy of the medical records. It has been suggested that genetic information, and particularly polygenic risk, is often predictive, particularly for highly heritable conditions. In this talk, I will describe an evaluation of methylation patterns (modifications to the DNA) as a predictive tool that can be used for EHR imputation. We show that methylation provides a better imputation performance when compared to genetic or EHR data. Our approach uses a new tensor deconvolution of bulk DNA methylation to obtain cell-type-specific methylation that is in turn used for imputation.
BIO: Eran Halperin is the SVP of AI and Machine Learning in OptumLabs (United HealthGroup), and a professor in the departments of Computer Science, Computational Medicine, Anesthesiology, and Human Genetics at UCLA. Prior to his current position, he held research and postdoctoral positions at the University of California, Berkeley, the International Computer Science Institute in Berkeley, Princeton University, and Tel-Aviv University. Dr. Halperin’s lab developed methods for a variety of health-related domains, including different genomic domains (genetics, methylation, microbiome, single-cell RNA), and medical applications (medical imaging, physiological waveforms, and electronic medical records). He published more than 150 peer-reviewed publications, and he received various honors for academic achievements, including the Rothschild Fellowship, the Technion-Juludan prize for technological contribution to medicine, the Krill prize, and he was elected as an International Society of Computational Biology (ISCB) fellow.
Irene Y. Chen
Clinical Machine Learning group at MIT's Computer Science and Artificial Intelligence Lab (CSAIL)
TITLE: Machine Learning for Equitable Healthcare
ABSTRACT: Advances in machine learning and the explosion of clinical data have demonstrated immense potential to fundamentally improve clinical care and deepen our understanding of human health. However, algorithms for medical interventions and scientific discovery in heterogeneous patient populations are particularly challenged by the complexities of healthcare data. Not only are clinical data noisy, missing, and irregularly sampled, but questions of equity and fairness also raise grave concerns and create additional computational challenges. In this talk, I present two approaches for leveraging machine learning towards equitable healthcare. First, I demonstrate how to adapt disease progression modeling to account for differences in access to care. Using a deep generative model, we can correct for patient misalignment in disease onset time to learn more clinically useful disease subtypes. Second, I examine how to address algorithmic bias in supervised learning for cost-based metrics discrimination. By decomposing discrimination into bias, variance, and noise components, I propose tailored actions for estimating and reducing each term of the total discrimination. The talk concludes with a discussion about how to rethink the entire machine learning pipeline with an ethical lens to building algorithms that serve the entire patient population.
BIO: Irene Chen is a PhD student in the Clinical Machine Learning group at MIT's Computer Science and Artificial Intelligence Lab (CSAIL), advised by David Sontag. Her work centers on machine learning methods for improving clinical care and making it more equitable, as well as auditing and addressing bias in algorithmic models. Her work has been published in machine learning conferences (NeurIPS, AAAI) and medical journals (Nature Medicine, Lancet Digital Health), and has been covered by media outlets including MIT Tech Review, NPR/WGBH, and Stat News. She has been named a Rising Star in EECS by University of California Berkeley, Harvard, and University of Maryland. Prior to her PhD, Irene received her AB/SM from Harvard and worked at Dropbox.
TITLE: Machine Learning for Equitable Healthcare
ABSTRACT: Advances in machine learning and the explosion of clinical data have demonstrated immense potential to fundamentally improve clinical care and deepen our understanding of human health. However, algorithms for medical interventions and scientific discovery in heterogeneous patient populations are particularly challenged by the complexities of healthcare data. Not only are clinical data noisy, missing, and irregularly sampled, but questions of equity and fairness also raise grave concerns and create additional computational challenges. In this talk, I present two approaches for leveraging machine learning towards equitable healthcare. First, I demonstrate how to adapt disease progression modeling to account for differences in access to care. Using a deep generative model, we can correct for patient misalignment in disease onset time to learn more clinically useful disease subtypes. Second, I examine how to address algorithmic bias in supervised learning for cost-based metrics discrimination. By decomposing discrimination into bias, variance, and noise components, I propose tailored actions for estimating and reducing each term of the total discrimination. The talk concludes with a discussion about how to rethink the entire machine learning pipeline with an ethical lens to building algorithms that serve the entire patient population.
BIO: Irene Chen is a PhD student in the Clinical Machine Learning group at MIT's Computer Science and Artificial Intelligence Lab (CSAIL), advised by David Sontag. Her work centers on machine learning methods for improving clinical care and making it more equitable, as well as auditing and addressing bias in algorithmic models. Her work has been published in machine learning conferences (NeurIPS, AAAI) and medical journals (Nature Medicine, Lancet Digital Health), and has been covered by media outlets including MIT Tech Review, NPR/WGBH, and Stat News. She has been named a Rising Star in EECS by University of California Berkeley, Harvard, and University of Maryland. Prior to her PhD, Irene received her AB/SM from Harvard and worked at Dropbox.
Michal Rosen-Zvi
Director, AI for Accelerated HC&LS Discovery IBM Research
TITLE: Acceleration of Biomarker Discovery in Multimodal Data of Cancer Patients - Promising Results
ABSTRACT: Artificial Intelligence (AI) technologies have recently demonstrated performance that matches radiologists' accuracy in a number of specific tasks, particularly in cancer detection in breast mammography images and chest CT and Xray. In recent years, AI solutions have shown to be capable of assisting radiologists and clinicians in detecting diseases, assessing severity, automatically localizing and quantifying disease features, or providing an automated assessment of disease prognosis. It remains an open question whether these technologies, when applied to multimodal data of cancer patients, can early detect the cancer type and hint to novel biomarkers associated with the disease. In this study I will discuss the advanced in developments of computational tools to assess cancer heterogeneity in general and will focus on AI for imaging in particular. Initial results that combine clinical and imaging data for assessing pathology will be shared. The talk is based on work performed with many collaborators who co-authored the following three papers: (i) Born, J., Beymer, D., Rajan, D., et al., 2021. On the role of artificial intelligence in medical imaging of COVID-19. Patterns, 2(6), p.100269. (ii) Kashyap, A., Rapsomaniki, M.A., Barros, V., et al., 2021. Quantification of tumor heterogeneity: from data acquisition to metric generation. Trends in Biotechnology, and (iii) Shoshan, Y., Bakalo, R., Gilboa-Solomon, F. et al. 2022. Artificial Intelligence for Reducing Workload in Breast Cancer Screening with Digital Breast Tomosynthesis. Radiology, p.211105.
BIO: Dr. Rosen-Zvi is the Director, AI for Accelerated Healthcare &Life Sciences Discovery at IBM Research and a visiting Professor at the Faculty of Medicine, the Hebrew University. She is also heading the AI for Healthcare Department at IBM Research, Haifa. Michal holds a PhD in computational physics and completed her postdoctoral studies at UC Berkeley, UC Irvine, and the Hebrew University in the area of Machine Learning. She joined IBM Research in 2005 and has since led various projects in the area of machine learning and healthcare and was recognized for her contribution e.g. to AI technologies in wafer production and contributions to partnerships with pharmaceutical companies such as Guerbet and Teva. Michal has published more than 40 peer-reviewed papers that were cited more than 5000 times according to Google Scholar. She is a member of IBM Industry Academy, a member of the Israeli National Council of Digital Health and Innovation and an elected member of the board of the Israeli Society for HealthTech.
TITLE: Acceleration of Biomarker Discovery in Multimodal Data of Cancer Patients - Promising Results
ABSTRACT: Artificial Intelligence (AI) technologies have recently demonstrated performance that matches radiologists' accuracy in a number of specific tasks, particularly in cancer detection in breast mammography images and chest CT and Xray. In recent years, AI solutions have shown to be capable of assisting radiologists and clinicians in detecting diseases, assessing severity, automatically localizing and quantifying disease features, or providing an automated assessment of disease prognosis. It remains an open question whether these technologies, when applied to multimodal data of cancer patients, can early detect the cancer type and hint to novel biomarkers associated with the disease. In this study I will discuss the advanced in developments of computational tools to assess cancer heterogeneity in general and will focus on AI for imaging in particular. Initial results that combine clinical and imaging data for assessing pathology will be shared. The talk is based on work performed with many collaborators who co-authored the following three papers: (i) Born, J., Beymer, D., Rajan, D., et al., 2021. On the role of artificial intelligence in medical imaging of COVID-19. Patterns, 2(6), p.100269. (ii) Kashyap, A., Rapsomaniki, M.A., Barros, V., et al., 2021. Quantification of tumor heterogeneity: from data acquisition to metric generation. Trends in Biotechnology, and (iii) Shoshan, Y., Bakalo, R., Gilboa-Solomon, F. et al. 2022. Artificial Intelligence for Reducing Workload in Breast Cancer Screening with Digital Breast Tomosynthesis. Radiology, p.211105.
BIO: Dr. Rosen-Zvi is the Director, AI for Accelerated Healthcare &Life Sciences Discovery at IBM Research and a visiting Professor at the Faculty of Medicine, the Hebrew University. She is also heading the AI for Healthcare Department at IBM Research, Haifa. Michal holds a PhD in computational physics and completed her postdoctoral studies at UC Berkeley, UC Irvine, and the Hebrew University in the area of Machine Learning. She joined IBM Research in 2005 and has since led various projects in the area of machine learning and healthcare and was recognized for her contribution e.g. to AI technologies in wafer production and contributions to partnerships with pharmaceutical companies such as Guerbet and Teva. Michal has published more than 40 peer-reviewed papers that were cited more than 5000 times according to Google Scholar. She is a member of IBM Industry Academy, a member of the Israeli National Council of Digital Health and Innovation and an elected member of the board of the Israeli Society for HealthTech.