Prof. Yuval Shahar


M.Sc committee - member

Department : Department of Software and Information Systems Engineering
Specialization in Computational Learning and BigData
Room : 228
בנין מערכות מידע וסייבר - 96
Phone : 972-74-7795120
Email :
Office hours :  

Office Hours

       בניין 96 חדר 226 בתיאום מראש Contact me

Phone Location

בנין מערכות מידע וסייבר - 962228

בנין מערכות מידע וסייבר - 962226


  • School, Location Major Subject, Degree, and Date
  • Hebrew University, Jerusalem, Israel, Natural Sciences, B.Sc., 1978
  • Hebrew University, Jerusalem, Israel, Medicine, M.D., 1981
  • IDF Computer Academy, Ramat Gan, Israel, System Analysis & Design, System Analyst, 1985
  • Bar-Ilan University, Ramat Gan, Israel , Mathematics & Computer Science, M.Sc. studies, 1985-1988
  • Yale University, New Haven, CT, Computer Science (A.I.), M.S., 1990
  • Stanford University, Stanford, CA, Medical Information Sciences, Ph.D., 1994

Research Interests

  • I am interested in artificial intelligence in general, and in its uses within medical domains in particular. More specifically, I am interested in temporal reasoning and planning in general, and in clinical decision-support applications in particular. I am also interested in general methods for knowledge representation and knowledge acquisition, and in general, in reusable and sharable, problem-solving methods. Apart from classical artificial intelligence techniques, I am also interested in applications of theoretical computer science techniques to such problems. I also have an interest in medical decision analysis.

Research Projects

  • A task common to many application domains is the analysis of data accumulated over time, leading to identification of past and present trends and to episodic decisions made on the basis of the previous and the current state of the world. An example of such a task in the medical domain is managing patients who are being treated with clinical guidelines. An inherent requirement of such tasks is to accumulate and to analyze patient data over time and constantly to revise an assessment of the patient's state by abstracting higher-level, context-sensitive concepts from the raw input data. These higher-level concepts can be used for summarizing large medical databases, for monitoring, for replanning therapy, for providing explanations to a user of a decision-support system, for (temporal) data mining and knowledge discovery, and as a basis for a more intelligent dialog between an automated decision-support system and a human health-care provider.
  • My work focuses on defining basic knowledge-based, domain-independent temporal-abstraction mechanisms and the formal knowledge needed to instantiate them in any particular medical domain. Formalization of temporal-abstraction knowledge supports the acquisition, representation, maintenance, reuse, and sharing of that knowledge. I have therefore defined a knowlede-based temporal-abstraction framework, implemented it as the R?SUM? system, and tested it in several clinical domains. The framework has been expanded and embedded in a larger architecture, Tzolkin, which combines temporal-reasoning and temporal-maintenance services. Tzolkin had been used within the EON component-based architecure for guideline-based care. An extension of Tzolkin is the IDAN temporal mediator at Ben Gurion University's Medical Informatics Research Center .
  • I am also leading the KNAVE (knowledge-based Navigation of Abstractions for Visualization and Explanation) project, which focuses on an interactive framework for visualization and exploration of time-oriented data (e.g., patient clinical data) and their multiple-levels of abstractions. the KNAVE project has been extended into the KNAVE-II project at Ben Gurion University's Medical Informatics Research Center , evaluated by Martins et al. (2008) at the Veterans Administration Palo Alto Health Care Center, and later extended into the VISITORS project for query , interactive visual exploration , and exploration of the temporal associations of the time-oriented data of multiple patients.
  • I also am interested in (therapy) plan generation, revision, recognition and critiquing in clinical domains; I have previously set up the Asgaard project, ongoing in several countries, which investigates these tasks. I am also leading the Digital Electronic Guideline Library (DeGeL) project, which creates a distributed framework for specification, retrieval, and use of clinical guidelines.
  • The DeGeL , IDAN , KNAVE-II, VISITORS , and other related projects are part of my Medical Informatics Research Center , a faculty-wide research center whose main laboratory is at the Department of Information Systems Engineering of Ben Gurion University of the Negev, Beer Sheva, Israel .
  • Finally, I am interested in decision-theoretical aspects of clinical decision making. I have previously led the PANDA project at Stanford University, which applied decision analytic methodologies to the domain of genetic consultation, taking into account the patient's characteristics and personal preferences, and the PANDEX project at BGU, which has implemented these methodologies on the WEB and investigated seevral methods for sensitivity analysis of the recommended optimal decision.