Segev Wasserkrug

I am a Senior Technical Staff Member at IBM Research - Haifa specializing in decision optimization and game theory. My passion centers around enabling organizations to get the most value from AI, especially through mathematical optimization and structured methods for decision-making. Right now, I’m focused on democratizing the use of mathematical optimization by making it easier and faster to deploy, and taking game theory science into the practical decision-making realm. In this way, I hope to promote better real-life choices in both single and multi-party settings based on the true constraints of the situation or environment—making it practical for use in enterprise decision-making.

My background includes experience as a senior analytics and optimization researcher. Since 2001, I have been researching advanced mathematical, analytical, and optimization techniques. Some of this involved applying advanced mathematical and analytical techniques to customer problems in a variety of domains including workforce optimization and scheduling, advanced water management, and  logistics. This has been based on a strong foundation in a variety of areas including artificial intelligence, optimization, complex event processing, computer science, software engineering, operations research, and simulation.

I received my Ph.D. in information systems engineering, along with my MSc and BSc in computer science from the Technion – Israel Institute of Technology. To date, I have co-authored over 35 academic publications and presentations and filed over 15 patents.

Current research focus

Mathematical optimization can provide significant benefits to making better decisions, but creating, maintaining and updating the required relevant models and solutions requires an enormous amount of time and effort, and a scarce skill set. Therefore, one of my current primary research focuses is leading work that uses AI to simplify the creation and lifecycle management process of decision optimization solutions; the idea is to radically reduce the time and skills required to create and maintain such solutions .

While significant benefits can be achieved by classical decision optimization, decision optimization typically involves making decisions in an environment that is oblivious to the decision-maker. Many real-world scenarios require making decisions in settings in which there are multiple self-interested parties, whose decisions can impact the benefits obtained by the other participants. Game theory is the mathematical science for analyzing such situations. However, it has significant gaps when it comes to providing prescriptive, computationally efficient solutions.  To address this, I am leading a team focused on advancing the core scientific basis of game theory and algorithmic game theory (for example, by coupling it with multi-agent reinforcement learning techniques) to enable better decision-making in real world situations—providing the best options given the actual constraints and real limitations of the environment.

Vision

I believes in the power of science and making decisions based on data and sound scientific principles. The  driving principle for my work is to enable more widespread use of science in making and implementing better decisions, in both single and multi-party settings.   I believe that decision optimization and  game theory can become much more fundamental in our lives by providing insights for the scientific outcomes of decisions being implemented.

Knowledge areas

Mathematical optimization, AI, reinforcement learning , game theory and algorithmic game theory, complex event processing   

Main research areas:

• Decision optimization, and particularly using AI to radically simplify the creation, maintenance and updating of decision optimization models
• Enabling practical decision making in multi-party settings by building upon and enhancing game theory, algorithmic game theory and multi agent reinforcement learning

Links

• My LinkedIn page
• Can AI solutions scale to meet their potential?
• A tale of 3 solutions - scaling AI solutions through the value loop pattern