Dr. Gonen Singer helps organizations make the right decisions

by combining machine learning algorithms and operations research models.

Over the past 12 years, before arriving at our Faculty, Dr. Singer has held two positions: the first, as a lecturer and later head of the Department of Industrial Engineering and Management at Afeka College of Engineering; the second, as co-founder of CB4, a company focused on patents in the field of pattern recognition. He started off as CTO, and after successful funding (resulting in additional centers in the UK and the US), was appointed head of activities in Israel. “I love the combination of academia and industry,” he says.

Dr. Singer (42), married+3, began his journey in academia in the academic reserve program. He completed his BSc and MSc prior to his mandatory service in the Air Force and then continued to complete his PhD during service. His PhD thesis focused on developing methods for optimal control in dynamic and stochastic environments. Upon leaving the military some 12 years ago, he was recruited to the Department of Industrial Engineering and Management at Afeka College of Engineering while it was still in its early days. A year later he had become head of the department.

While teaching at Afkea, he also founded CB4 together with his MSc supervisor, Prof. Irad Ben-Gal. The company offers automatic identification of operational problems by studying shopper behavior in retail chains. “CB4 started out from one research at Tel Aviv University. The company has enabled me to realize one of my greatest professional aspirations: turning academic research into an applied tool that can be used by customers worldwide, on a daily basis,” recounts Singer.

The company started out as a bootstrap venture, without funding, and only began raising funds some 5 years ago, when it sought to enter the European and US markets. Today, the company has 70 employees in three centers spread over three continents: Herzliya, Israel; London, UK; and New York, USA. Among its clients are Superpharm, Shufersal, Levi’s (US and China), and Media Market in Germany. “We’ve now become a well-established company with a strong executive infrastructure,” explains Singer, “so I decided it was the right time to leave Afeka, which emphasizes teaching, in favor of a university that emphasizes research, which I love.”

Singer, joined the Faculty of Engineering this past March and is currently helping Dr. Itzik Cohen to establish the Industrial Engineering and Information Systems program. The program was approved by the university, with the hopes of launching in the upcoming academic year, pending the approval of The Council for Higher Education. At the same time, the two researchers are starting the Process-Mining and Modeling laboratory, where they incorporate mathematical models in operations research with machine learning algorithms for the purpose of analyzing complex processes. “We’re seeking curious and creative MSc and PhD student with a strong mathematical background and programming skills.

Dr. Singer’s studies focus on problems that are real and mathematically complex, on the one hand, yet data-rich on the other, and that often cannot be solved using classic models. “One of my recent studies, for example, deals with assigning different roles in organizations to potential candidates. Mathematically speaking, we can develop a model that finds an optimal solution for this kind of assignment, according to organizational demands. In addition, the organization acquires data concerning candidates that were selected – years of formal education, time spent in prior roles, location, etc., as well as an indication of their level of success in their roles. The first stage of my research combined between interpretable machine learning models, meant to study a candidate’s chances of succeeding at a certain role while identifying the reasons for it. At the second stage, we incorporated the insights gained from machine learning algorithms as input for the operations research model, in an attempt to assign candidates while minimizing their dropout rates, and meeting the requirements of their role as defined by the organization.”

“I also engage in developing machine learning algorithms for ordinal classification problems, where the order has meaning for the goal values. In that context, I recently applied for a research grant where I intend to develop an algorithm for predicting the contagion rate of COVID-19, on a granular level of geography and time, based on parameters of weather and public’s movement patterns. Preliminary results were applied to 20 areas in Italy, and produced 90% precision in predicting the contagion rate 6 days in advance.”