Multi-Agent Artificial Intelligence - Artificial Intelligence Group

  • Postal addressShow on map

    Computer Science department, Science Faculty | Bar Ilan University, Room 225 / Building 216

    5290002 Ramat Gan


Organization profile

Organisation profile

The computer science department at Bar Ilan University, is a home to Israel's largest artificial intelligence group. The group is well-known worldwide for its many exceptional scientific achievements for the past two decades. It includes 11 senior faculty members and dozens of graduate students, post-docs, programmers and computer engineers. 
One of the main strengths of the group is intelligent robotics, and in particular the study of multi-robot systems, led by Prof. Gal Kaminka and Dr. Noa Agmon. Among the many interesting topics researched by the two, one can find problems of multi-robot planning in adversarial environments, multi-robot patrolling, navigation, coverage, task allocation and formation. Much of the group's work on multi-robot patrolling, exploration, and formations was highly publicized, and resulted in several technology transfer and patent programs. Agmon's focus is primarily on strategic behavior of robots in adversarial environments, and using theoretical means to model, analyze and solve those and other realistic robotic problems. Kaminka is more into the control of teams of robots and agents on the one hand, and its complementary monitoring by the agents themselves on the other.  He initiated the annual ARMS (Autonomous Robots and Multi-Robot Systems) workshop at AAMAS, which is now the main meeting point for all roboticists from the AAMAS community. One seminal result of his research is the realization that teamwork, as a set of general mechanisms for collaboration, can be automated and computerized. This makes it possible for robots to collaborate well, cheaply, and do so also with humans. 
Another area that attracts much of the group's attention is the study of the theoretical foundations of interactions among multiple software agents, with a focus on settings were the agents represent disparate entities, with conflicting interests. Using mathematical and algorithmic tools, Prof. Yonatan Aumann analyzes such settings, and develops methods to promote social goals, such as fairness and welfare. In an interesting recent project, Aumann and colleagues are extending the decades old study of so called ``cake cutting'' from the one dimensional case to two dimensions. In doing so, they have devised the first ever efficient protocols for fair-division of land. Various other aspects of multi-agent systems are studied at the AIM lab headed by Prof. Sarit Kraus. Kraus is recognized for her many contributions to AI subfields such as strategic negotiation, collaborative planning, human-agent interaction, coalition formation and non-monotonic reasoning. Nowadays she is involved, together with many partners both from industry and academia, in developing systems and technologies that will revolutionize various traditional professions. These include a virtual speech therapist, culture-sensitive systems that collaborate, negotiate and argue proficiently with people, systems for training law enforcement officials to interview witnesses and suspects and persuasion systems that generate advice for drivers regarding various different decisions involving conflicting goals. Other active members in the area of multi-agent systems are Dr. Avinatan Hassidim and Prof. David Sarne. Hassidim works in algorithmic mechanism design, developing new algorithms and implementing them in real world markets. His list of recent projects includes impressive implementations of new methods for the Israeli medical intern lottery, university admission to psychology and the design of a national school choice system. Sarne is focused in studying the role that information plays in multi-agent settings and in particular the dynamics resulting from introducing information brokers/platforms into such environments. Another topic of interest of his is the design of methods for intelligent information provisioning in human-agent interaction and in particular methods for intelligent advice provisioning.
The Bar-Ilan group is also very active in the area of natural language processing (NLP). The leading group members in this area are Professor Ido Dagan and Dr. Yoav Goldberg. Dagan and his group focus on applied semantic processing of texts, taking an empirical approach largely based on unsupervised and supervised learning. His current work involves developing a generic framework for representing the knowledge expressed in texts. This framework is based on using natural language constructions (rather than synthetic languages) as the basic building blocks for knowledge representation, aiming at an open and scalable scheme for representing unrestricted textual knowledge. While Dagan's work focuses on higher level semantic tasks, Goldberg's research involves the lower-level building blocks of natural language processing: analyzing sentences to their syntactic structures, deriving continuous vector representations for words, and building tools and methodologies that allow for cross-domain and cross-language generalizations.  Currently, Goldberg's group focuses on applications of advanced deep-learning techniques (recurrent and recursive neural networks) to structured prediction problems in natural language processing.
Machine learning research goes way beyond NLP at the AI group at Bar-Ilan University, with the lead of Prof. Moshe Koppel And Dr. Joseph Keshet. Koppel conducts research on a variety of machine learning applications including text categorization, image processing, speaker recognition and automated game playing. He is best known for his contributions to the branch of text categorization concerned with authorship attribution. In a string of papers, Koppel and colleagues solved many of the main problems in authorship, including authorship verification and authorship attribution with huge open candidate sets. Keshet focuses in general machine learning algorithms that are usually applied to speech, language and medical related problems. Among his recent projects, a system for automatic analysis of doppler echocardiography, algorithms for extracting of phonetic parameters from elicited speech with human-like accuracy, computerized analysis of speech pathologies, such as vocal cords paralyses and vocal cord lesions and spoken language vocal synchrony in psychotherapy sessions.
The AI group conducts also research in areas such as image processing, pattern recognition, and computational intelligence. Heading this effort is Prof. Nathan Netanyahu, who is known for his highly cited works on nearest neighbor searching, k-means clustering, and point pattern matching. In recent years, Netanyahu (along with his former Ph.D. student, Dr. Omid David) has focused on developing state-of-the-art methods, based on evolutionary computation/deep learning, for a variety of tasks, including object/painter classification, stereo matching, image registration, jigsaw puzzle solving, computer chess, and malware classification.
Being Israel's largest AI center, the group organizes and hosts BISFAI, the biennial regional artificial intelligence symposium (the largest AI event in the middle east!), which is a warm, friendly event that takes place for more than 25 years now. World renowned AI researchers regularly visit and lecture at BISFAI, which gives an opportunity for the Israeli AI community, especially the students, to meet their peers from abroad. To many, BISFAI has become the premier event of AI happening in Israel.


Dive into the research topics where Multi-Agent Artificial Intelligence - Artificial Intelligence Group is active. These topic labels come from the works of this organization's members. Together they form a unique fingerprint.