Rational design of drug delivery systems for selective cancer therapy

Ofra Benny, Ph.D, Associate Professor

Institute for Drug Research, The School of Pharmacy,  Faculty of Medicine, Campus Ein Karem,

The Hebrew University of Jerusalem

Understanding the complexity of tumor biology and the underlining cellular processes involved in disease progression, emphasizes the need to identify varying cell behaviors that can be integrated to reveal clinical relevant information. The effort to detect molecular biomarkers for targeted therapy are limited by the dynamic nature of cellular molecular pattern and their heterogeneity. Measuring mechanical properties of cancer cells was suggested as a complimentary approach that can detect malignant features of cancer cells. Here we provide a comprehensive study that provide a trio correlation between cell mechanics, their tendency to interact with particles in a phagocytic-like manner, and their aggressiveness. By bi-directional study we found that the capacity for particle uptake is an imprinted cell behavior and “phagocytic” cancer cell population can be enriched over time, and correlate with higher malignancy and greater deformability and adhesion as demonstrated in vitro and in vivo. In agreement, greater malignancy was correlated with higher particle uptake and less rigid phenotype. Our study provides new mechanistic insights related to the role of mechanical tuning during cancer progression, and, importantly to suggest a future direction of using particle uptake as diagnostic tool for precision medicine.

ההרצאה תתקיים ביום ראשון 08.12.2019, בשעה 14:00

בחדר 315, הבניין הרב תחומי, אוניברסיטת תל אביב

(The talk will be given in English)

Speaker:     Dr. Alon Gonen
                     University of California San Diego

Wednesday, December 18^th, 2019
15:00 - 16:00

Room 011, Kitot Bldg., Faculty of Engineering

Beyond Worst Case In Machine Learning: The Oracle Model

Abstract

In recent years there has been an increasing gap between the success of machine learning algorithms and our ability to explain their success theoretically. 

Namely, many of the problems that are solved to a satisfactory degree of precision are computationally hard in the worst case. Fortunately, there are often reasonable assumptions which help us to get around these worst-case impediments and allow us to rigorously analyze heuristics that are used in practice. 

In this talk I will advocate a complementary approach, where instead of explicitly characterizing some desired "niceness" properties of the data, we assume access to an optimization oracle that solves a relatively simpler task. This allows us to identify the sources of hardness and extend our theoretical understanding to new domains. Furthermore we will show that seemingly innocents (and arguably justifiable) modifications to the oracle can lead to tractable reductions and even to bypass hardness results. 

We demonstrate these ideas using the following results: i) An efficient algorithm for non-convex online learning using an optimization oracle. b) A faster boosting algorithm using a "simple" weak learner. iii) An efficient reduction from online to private learning.

Joint works with Naman Agarwal, Noga Alon, Elad Hazan, and Shay Moran.

Short Bio
Alon Gonen is a postdoctoral at University of California San Diego. Prior to that he was a postdoctoral at Princeton university. He completed his MSc and PhD at the Hebrew University in under the supervision of Shai Shalev-Shwartz. His research focuses on theoretical aspects of machine learning. 

 (The talk will be given in English)

Speaker:     Dr. Vardan Papyan
                    Department of Statistics at Stanford University

Monday, December 16^th, 2019
15:00 - 16:00

Room 011, Kitot Bldg., Faculty of Engineering

Traces of Class/Cross-Class Structure Pervade Deep Learning Spectra

Abstract

Numerous researchers recently applied empirical spectral analysis to the study of modern deep learning classifiers, observing spectral outliers and small but distinct bumps often seen beyond the edge of a “main bulk”. This talk presents an important formal class/cross-class structure and shows how it lies at the origin of these visually striking patterns. The structure is shown to permeate the spectra of deepnet features, backpropagated errors, gradients, weights, Fisher Information matrix and Hessian, whether these are considered in the context of an individual layer or the concatenation of them all. The significance of the structure is illustrated by (i) proposing a correction to KFAC, a well known second-order optimization algorithm for training deepnets, and (ii) proving in the context of multinomial logistic regression that the ratio of outliers to bulk in the spectrum of the Fisher information matrix is predictive of misclassification.

Short Bio
Vardan Papyan received his B.Sc. degree in 2013 from the Computer Science Department at the Technion - Israel Institute of Technology, Haifa, Israel. He later received his M.Sc. and Ph.D. in 2017 from the same department, under the supervision of Prof. Michael Elad. He is currently a postdoctoral researcher, advised by David Donoho, at the Department of Statistics at Stanford University. His research interests include machine learning and more specifically deep learning, data science, signal and image processing, and high-dimensional statistics.‏

הדר וייס, מנכ"ל ומייסד חברת פיר5 יספר את סיפור החברה שעברה מתל אביב לקליפורניה ואיך דווקא משם פיר5 הגיעה למדינות רחוקות כמו ברזיל, ארגנטינה, וייטנאם, תאילנד והודו. עוד יספר על מה שנלמד בדרך - כוחם של הפשטות, הצניעות וקשר אנושי ישיר. מה הטריקים שעבדו לו ולשותפיו ומה היו הקשיים, כולל הקושי לעבור ממיינדסט של מהנדס שכיר למחשבה עסקית.

הדר וייס בעל תואר ראשון במדעי המחשב מאוניברסיטת תל אביב, בו הכיר את שני שותפיו. מתכנת מגיל צעיר ועבד כמפתח כ-15 שנה. את פיר5 הקים ב-2012. גר בסן פרנסיסקו שנתיים והשתתף בתוכנית Y Combinator.  שב לגור בתל אביב ב-2018. 

חצי שעה ראשונה התכנסות. ההשתפות בהרשמה מראש בלבד.

בוגרים, סטודנטים וסגל הפקולטה להנדסה באוניברסיטת תל אביב ללא תשלום - להרשמה

אורחים בתשלום סמלי - לתשלום