Estimating Information-Theoretic Measures with Classification Algorithms

Yuval Shalev

  Ph.D candidate at the Department of Industrial Engineering in the Tel-Aviv University, 

  under the guidance of Prof. Irad Ben-Gal and Dr. Amichai Painsky.
 
https://us02web.zoom.us/j/88556814331?pwd=REJvREpGNzVJUlQ4Y0lRSFczTHJkUT09
 
   

Abstract:
 
Estimating the information-theoretic measures of discrete random variables is a fundamental problem in information theory and related fields. This problem has many applications in various domains, including machine learning, statistics and data compression. Over the years, a variety of estimation schemes have been suggested. However, despite significant progress, most methods still struggle when alphabet size is large, random variables are from high dimensional space or when time dimension is involved. In this work, we introduce a holistic solution for this problem that is based on classifiers models. A set of these models is trained on the joint distribution of the random variables ,to obtain estimators for fundamental measures such as entropy, mutual information and transfer entropy. We focus on estimating these measures in scenarios where other methods usually fail. We  provide an experimental study of several use cases that demonstrates the advantages of our proposed scheme over state-of-the-art methods.
 
Bio:
 
Yuval Shalev is a Ph.D candidate at the Department of Industrial Engineering in the Tel-Aviv University,   under the guidance of Prof. Irad Ben-Gal and Dr. Amichai Painsky. Previously, he completed his M.Sc studies in physics in the Hebrew University. His research interests include Machine learning, deep learning and their connection to information theory. In parallel to its academic studies, Yuval has been working in the financial industry since 2008 as a quant and a data scientist.  Today, He leads the research team in the data science group of Citi innovation lab in Israel. 

"ZOOM" SEMINAR

SCHOOL OF MECHANICAL ENGINEERING SEMINAR
Monday, January 25, 2021 at 14:00

Tristability, simulations mode actuation, and limit point behaviour in initially curved coupled micro beams

Dr. Lior Medina

Nanoscience Centre, Department of Engineering, University of Cambridge, CB3 0FF, United Kingdom

lm872@cam.ac.uk

Curved structures have been regarded as a benchmark problem for the study of nonlinear structures, presenting a myriad of complex responses, even more so when under the influence of nonlinear, displacement-dependent, electrostatic load. Recently, the focus has gravitated to two-dimensional structures such as plates and shells, as well as Kirigami based structures, which added to the overall complexity of their analysis, theoretically as well as experimentally, while ensuing investigations as to their intrinsic stability behaviour and potential applications. With the entrance of coupled structures into the fold, new insights have offered a glimpse to uncharted territories, heretofore unencountered, where nonlinear phenomena are inherently pronounced or alternately, controlled, while presenting complex behaviour that can even match two-dimensional structures. Among such phenomena is multi-stability, emanating from the coupling of two bistable constructs, forming a complex metastructure. Indeed, multi-stability, which has more than two stable equilibria, has been a highly sought-after property for a long time. The incentive for their study lies in their potential, exemplified through applications which benefit from an increased number of values. Among such applications it is possible to name non-volatile mechanical memoirs (NVMM), where a multi-valued element can lead to higher storage density, allowing to store more date due to increased number of bits. However, current studies have either produced structures too cumbersome for fabrication on the one hand, or too large (in the centimetre scale) on the other hand, making it difficult to realise multi-stable applications in the microscale. In the current study, a comprehensive and rigorous analysis is carried for a double beam metastrcutre via reduced order (RO) modelling, arc-length “Riks” continuation method, and multi degree-of-freedom (DOF) dynamics, to characterise and disclose capabilities embodied in such a structure, beyond bistability.

On the whole, it is shown that while such a structure may seem to be innocuous and simplistic at first glance, it holds a vast range of intrinsic responses, hitherto unknown, from parallel mode actuation to several forms of tristability. By acquiring the ability to transform such a structure to become bistable or tristable, depending on defining parameters, a whole new frontier of multistable based applications can arise on the microscale, while also providing a model that can be used by engineers in the early design stages. In addition, when taking into account the different multistable structures that were investigated up until now, it is also evident that the present study brings forth a relatively compact structure, promoting the miniaturisation of various applications, such as a three bit NVMM, multi valued logical gate, or multi-acceleration threshold sensors.

https://zoom.us/j/96584758181?pwd=WC9PMXdsYzJ3NFdEN2Q5ZUtOZEVjdz09