(The talk will be given in English)

Speaker:     Dr. Ran Gilad-Bachrach
                              Microsoft 

Monday, May 7^th, 2018
15:00 - 16:00

Room 011, Kitot Bldg., Faculty of Engineering

Private Collaborative Learning of Intelligible Models on Vertically Split Data

Abstract

A common privacy scenario in machine learning is one in which possibly several different parties hold fragments of a training dataset, and want to train a model together without revealing any unnecessary information to each other about their data. We present methods for training intelligible models, namely Generalized Additive Models (GAMs), on datasets that are vertically split across multiple parties. Our protocol uses fully homomorphic encryption to guarantee data privacy, and is proved to be secure in the semi-honest security model under mild non-collusion assumptions. We prove the privacy properties of our solution as well as demonstrate its practicality in empirical evaluation.

Bio

Ran Gilad-Bachrach is a researcher at Microsoft Research. He studies multiple aspects of machine learning, currently he focuses on Private-AI and Precision-Psychology. Private-AI is the study of privacy preserving and collaboration enabling machine learning. Precision-Psychology is the study of the applications of data science to improving outcomes in helping people going through behavioral changes or fighting disorders. Ran earned his Ph.D. from the Hebrew University of Jerusalem and worked at Intel Research and Bing before joining Microsoft Research.

School of Mechanical Engineering Seminar
Monday, May 7, 2018 at 14:00
Wolfson Building of Mechanical Engineering, Room 206

Solitary waves on ferrofluids jet

Prof. Emilian Parau

 Univ. of East Anglia UK

'Travelling axisymmetric waves on the surface of a cylindrical ferrofluid jet are investigated. An azimuthal magnetic field is generated by an electric current flowing along a stationary metal rod which is mounted along the axis of the moving jet. Solitary waves are computed and their time evolution is discussed. Comparisons with weakly nonlinear theories such as KdV, and experiments are presented.'

School of Mechanical Engineering Seminar
Monday, June 18, 2018 at 14:00
Wolfson Building of Mechanical Engineering, Room 206

A sublaminate meso-mechanics damage model for low-velocity impact analysis of multi-layered laminates

Yarden Zaki

MSc Student of Prof. Rami Haj-Ali

Low velocity impact (LVI) can induce hidden damage in laminated composite structures.  Thus this is a major risk source that may reduce strength and stiffness and limit serviceability. Wide spread delaminations are frequent outcome of LVI.   Hence, it is essential to develop predictive mechanical models for LVI damage.

 In this study, a through-thickness meso-mechanical sublaminate homogenization model is proposed for the damage analysis of composite plates subjected to low-velocity impact (LVI), see Figure-1. The sublaminate model performs through-thickness homogenization using the smallest repeated stack sequence of the multi-layered laminate. In-plane and out-of-plane damage models are introduced and examined for their ability to initiate and propagate progressive damage in the laminate. The sublaminate includes a thin cohesive layer that represents the intra-layer delamination damage mode.  Out-of-plane normal and shear failure criteria are combined to detect cohesive damage. The proposed sublaminate damage framework is implemented as an external user material of theABAQUS/explicit FEA program. LVI analyses of composite plates are performed and the resulted impact damage shapes are compared with test results taken from the literature.  Good prediction ability is shown for impact of multi-layered plates having several repeated stacks through their thickness.

Bio-Composites Based on Unique Coral Collagen Fibers towards

Tissue Engineered Blood Vessels

Shir Wertheimer

M.Sc. student of Prof. Rami Haj-Ali

Cardiovascular Diseases (CVDs), e.g. peripheral arterial diseases, cerebrovascular disease, and particularly coronary artery occlusion, constitute the leading cause of death worldwide. Approximately 31% of all global deaths are caused by CVDs, of which 42% are attributable to coronary artery disease. These are manifested by, inter alia, the mechanical degradation of the blood vessels.  The mechanical strength of the blood vessel is provided primarily by type I collagen fibers which are mainly found in the outer layer of blood vessels, known as the adventitia.

This study introduces a new bio-composite material system consisting of unique and long collagen fibers derived from corals – embedded within an alginate hydrogel matrix. The new bio-composite layers were used to fabricate a graft to be used towards tissue engineered blood vessels.  These constructs consisted of both circumferentially and longitudinally oriented collagen fibers. Mechanical properties of the grafts such as compliance and hoop stress-strain curve were investigated via a new experimental setup constructed in our lab for this purpose, allowing applied internal pressure levels (0-300 mmHg) with measured external deformations using an optical extensometer. Furthermore, numerical finite element simulations of the graft predicted close stiffness values to the measured, especially for relatively high pressure loads.  

Next, a biological biocompatibility study was conducted to examine cell growth within the manufactured composite construct.  Towards that goal, fibroblast cells were seeded within the collagen fibers alone and monitored in order to generate quantitative growth metrics. Subsequently, different bio-composite collagen-reinforced hydrogels were also fabricated and cell-seeded for similar bio-compatibility studies. The cells in all samples used for the biocompatibility tests were alive and well-functioning for more than 28 days.  They demonstrated higher growth rates during the first two weeks, followed by lower yet generally positive growth rates for the following two weeks. The cell orientation was shown to follow that of the collagen fibers for the duration of the experiment.

The novelty of this study is manifested in the use of naturally derived long collagen fibers for the development of a new class of tissue-engineered grafts. The proposed novel bio-composite and associated construct show a great potential for future tissue engineered replacements of blood vessels.