-My Research

Abstract—This paper considers the problem of self-interested
agents engaged in costly exploration when individual findings
benefit all agents. The purpose of the exploration is to reason
about the nature and value of the different opportunities available
to the agents whenever such information is a priori unknown.
While the problem has been considered for the case where the
goal is to maximize the overall expected benefit, the focus of
this paper is on settings where the agents are self-interested,
i,e, each agent’s goal is to maximize its individual expected
benefit. The paper presents an equilibrium analysis of the model,
considering both mixed and pure equilibria. The analysis is
used to demonstrate two somehow non-intuitive properties of
the equilibrium cooperative exploration strategies used by the
agents and their resulting expected payoffs: (a) when using mixed
equilibrium strategies, the agents might lose due to having more
potential opportunities available for them in their environment;
and (b) if the agents can have additional agents join them in the
exploration they might prefer the less competent ones to join the

In many domains, an autonomous agent needs to reliably predict the distribution of behaviors of a population rather than the behavior of a single agent. For example, when playing the ultimatum game against several unknown opponents from a large known population, the agent can perform better by extracting its best-response strategy based on the distribution of the acceptance value in that population. In this paper, we demonstrate the efficacy of Peer-Designed-Agents (PDAs) for producing a distribution of behaviors that highly resembles the distribution of actual behaviors of a specific population of interest. This is obtained through extensive experiments with more than 700 different individuals and 132 PDAs, using eight game variants from three different domains and two different statistical tests.  The analysis of the results demonstrates that PDAs' technology is an effective means for generating a reliable distribution of behaviors of a population of interest, as long as the similarity between the group of PDAs' developers and the latter population is sufficiently high.  Moreover, a comprehensive comparison with the results of Elicited-Strategy-Agents (ESAs) shows that there is much more to PDA technology than simply an expression of strategy.

Joint Search with Self-Interested Agents and the Failure of Cooperation Enhancers - AIJ 15

This paper considers the problem of autonomous agents that need to pick one of several options, all plausible however differ in their value, which is a priori uncertain and can be revealed for a cost. The agents thus need to weigh the benefits of revealing further values against the associated costs. The problem is often referred to as "economic search" or "optimal stopping". This paper addresses the problem in its multi-agent joint form, such that not a single but rather a group of agents may benefit from the fruits of the search. The paper formally introduces and analyzes the joint search problem, when carried out fully distributedly, and determines the strategies to be used by the agents both when fully cooperative and when self-interested. For the self-interested case the analysis is based on equilibrium considerations, considering Bayesian Nash equilibria. The analysis is used to demonstrate that elements that can easily be proved to be beneficial with fully cooperative agents' search (e.g., extension of the search horizon, increase in the number of cooperating agents, improvement in agents' search competence or their search fallback and the use of continuous communication along the search) can actually degrade individual and overall expected utility in the self-interested case. The analysis contributes both to the advancement of joint search theories, and offers important insights for system designers, enabling them to determine the mechanisms that should be included in the markets and systems they design, whenever costly-search is an inherent feature of their market.


A lecturer for the course- Databases Design & Management at Tel-Aviv Yafo College

A teaching assistant for the courses- Object-Oriented/C++, Introduction to Computer Science, Statistic Methods for Computer Science and Machine Learning at Bar-Ilan University

Mathematics Teacher for students with learning disabilities at Nitzan - Institute for students with learning disabilities

A teaching assistant for the courses Introduction to Computer Science, Introduction to System Programming and Object-Oriented/C++ at Shenakr

A teaching assistant for the course- Object-Oriented/C++ at the College of Management

A member of Smart society project