פרופ' גדעון צבי שגב

Regulation of cerebral blood flow and metabolism

מרצה אורח

דר' מיכאל נ. דירינג'ר

Dr Michael N Diringer

The brain is a highly metabolic organ.  Oxidative metabolism of 2-carbon fragments in mitochondria is the primary source of energy using derived from oxygen and glucose.  The brain stores no substrates; glucose is actively transported by carrier-mediated facilitated diffusion, whereas oxygen passively diffuses down its concentration gradient.  While glucose supply is abundant, the brain utilizes 30-50% of delivered oxygen.  Cerebral oxygen delivery is the product of cerebral blood flow (CBF) and blood oxygen content (CaO2).

CBF is tightly coupled to metabolic demand. Regulatory factors include oxygen, potassium, adenosine, lactic acid.  CBF is also strongly influenced by CO2 and perfusion pressure.  It is the complex interplay all these factors that ultimately determines CBF.

The initial response to a fall in perfusion pressure is vasodilation in order to maintain stable CBF (pressure autoregulation).  If flow falls further brain oxygen extraction rises approaching 100%.

Dr. Michael N. Diringer is a Professor of Neurology, Neurological Surgery, Anesthesiology, and Occupational Therapy at Washington University School of Medicine, St. Louis, Missouri. He is also the Chair of the Board of Directors at Mid-American Transplant and the Editor-in-Chief of Neurocritical Care. Dr. Diringer is the author or 200 peer-reviewed papers and over 50 book chapters and invited publications.

(The talk will be given in English)

Speaker:     Dr. Yaniv Romano
                     Department of Statistics at Stanford University

Wednesday, January 15^th, 2020
15:00 - 16:00

Room 011, Kitot Bldg., Faculty of Engineering

Reliability, Equity, and Reproducibility in Modern Machine Learning 

Abstract

Modern machine learning algorithms have achieved remarkable performance in a myriad of applications, and are increasingly used to make impactful decisions in the hiring process, criminal sentencing, healthcare diagnostics and even to make new scientific discoveries. The use of data-driven algorithms in high-stakes applications is exciting yet alarming: these methods are extremely complex, often brittle, notoriously hard to analyze and interpret. Naturally, concerns have raised about the reliability, fairness, and reproducibility of the output of such algorithms. This talk introduces statistical tools that can be wrapped around any “black-box" algorithm to provide valid inferential results while taking advantage of their impressive performance. We present novel developments in conformal prediction and quantile regression, which rigorously guarantee the reliability of complex predictive models, and show how these methodologies can be used to treat individuals equitably. Next, we focus on reproducibility and introduce an operational selective inference tool that builds upon the knockoff framework and leverages recent progress in deep generative models. This methodology allows for reliable identification of a subset of important features that is likely to explain a phenomenon under-study in a challenging setting where the data distribution is unknown, e.g., mutations that are truly linked to changes in drug resistance.

Short Bio
Yaniv Romano is a postdoctoral scholar in the Department of Statistics at Stanford University, advised by Prof. Emmanuel Candes. He earned his Ph.D. and M.Sc. degrees in 2017 from the Department of Electrical Engineering at the Technion—Israel Institute of Technology, under the supervision of Prof. Michael Elad. Before that, in 2012, Yaniv received his B.Sc. from the same department. His research spans the theory and practice of selective inference, sparse approximation, machine learning, data science, and signal and image modeling. His goal is to advance the theory and practice of modern machine learning, as well as to develop statistical tools that can be wrapped around any data-driven algorithm to provide valid inferential results. Yaniv is also interested in image recovery problems: the super-resolution technology he invented together with Dr. Peyman Milanfar is being used in Google's flagship products (Pixel 2/XL Phones, Google Clips, Google+, and Motion Stills), increasing the quality of billions of images and bringing significant bandwidth savings. In 2017, he constructed with Prof. Michael Elad a Massive Open Online Course (MOOC) on the theory and practice of sparse representations, under the edX platform. Yaniv is a recipient of the 2015 Zeff Fellowship, the 2017 Andrew and Erna Finci Viterbi Fellowship, the 2017 Irwin and Joan Jacobs Fellowship, the 2018–2020 Zuckerman Postdoctoral Fellowship, the 2018–2020 ISEF Postdoctoral Fellowship, the 2018–2020 Viterbi Fellowship for nurturing future faculty members, Technion, and the 2019–2020 Koret Postdoctoral Scholarship, Stanford University.‏

Speaker: Ophir Yaniv

M.Sc. student under the supervision of Prof. Nahum Kiryati and Dr. Arnaldo Mayer

Sunday, January 26^th, 2020 at 15:30

Room 011, Kitot Bldg., Faculty of Engineering

Towards Automatic Lesion Detection in 3D Prostate MRI Scans

Abstract

Prostate cancer is the second most common cancer among men. Early detection is critical to the success of its treatment. Thanks to the excellent contrast it provides in soft tissues, prostate MRI has become the tool of choice in prostate cancer imaging.

We present a novel approach for the automated detection of prostate lesions in 3D MRI scans. Our methodology divides the workflow into two stages: 1) Segmentation of the prostate within the MRI scan to remove unrelated tissues. 2) Detection of suspected malignancies within the segmented prostate.

Previous methods of 3D prostate segmentation required high computational power and memory which are usually not available on the PCs used by radiologists. We propose a novel 3D deep neural network architecture, called V-net Light (VnL), that is based on a computationally efficient 3D Module, called 3D Light. The resulting network minimizes the number of parameters while maintaining state-of-the-art segmentation results.

Qualitative and quantitative validation of the proposed architecture will be presented.

To conclude, several directions for the lesion detection stage, to be developed in continuation of this research, will be discussed.

 (The talk will be given in English)

Speaker:     Dr. Oren Barkan
                    Microsoft, Israel

SUNDAY, January 12^th, 2020
15:00 - 16:00

Room 011, Kitot Bldg., Faculty of Engineering

VBNets: Learning Entity Representations via Variational Bayesian Networks

Abstract

Learning entity representations is an active research field. In the last decade, both the NLP and recommender systems communities introduced a plethora of methods for mapping words, items and users to vectors in a latent space. The vast majority of these works utilize implicit co-occurrences relations (e.g.  co-occurrences of words in text, co-consumption of items by users) for learning the latent entity vectors. Yet, often, additional side information in the form of explicit (e.g. hierarchical, semantic, syntactic) relations can be leveraged for learning finer embeddings. 

In this talk, we present Variational Bayesian Networks (VBNets) - A novel scalable hierarchical Bayesian model that utilizes both implicit and explicit relations for learning entity representations. VBNets are designed for Microsoft Store and Xbox services that handle around a billion users worldwide. Different from point estimate solutions that map entities to vectors and are usually over confident, VBNets map entities to densities in the latent space and hence model uncertainty. VBNets are based on analytical approximations of the intractable entities' posterior and the posterior predictive distribution of the data. We demonstrate the effectiveness of VBNets on linguistic, recommendations, and medical informatics tasks, where it is shown to outperform other alternative methods that facilitate Bayesian modeling with or without semantic priors. In addition, we show that VBNets produce superior representations for rare words and cold items. If time permits, we will give a brief overview of several recent deep learning works in the domains of deep neural attention mechanisms, multiview representation learning and inverse problems with applications for natural language understanding, recommender systems, computer vision, sound synthesis and biometrics. 

Short Bio
Oren Barkan is a Principal Researcher at Microsoft, where he was previously a post-doctoral researcher, collaborating with Microsoft Research UK and Microsoft Israel. Prior to that, he was with Google Research and IBM Research. He received his Ph.D. from Tel Aviv University, under the supervision of Prof. Amir Averbuch. His research interests are deep neural attention mechanisms, representation learning, multiview learning, Bayesian inference and inverse problems with applications for computer vision, natural language understanding, recommender systems, speech analysis, sound synthesis, biometrics, inflation forecasting, healthcare and medical informatics. He is the author of more than 40 research papers and patents.‏