SoCS 2017
Invited Speakers

Important Links

Colocated Conference

Past SoCS Conferences
2016 Tarrytown, New York, USA
2015 Ein Gedi, Isral
2014 Prague, Czech Republic
2013 Leavenworth, WA
2012 Niagara Falls, Canada
2011 Barcelona, Spain
2010 Atlanta, GA
2009 Lake Arrowhead, CA
2008 Chicago, IL

SoCS 2017: The 10th Annual Symposium on Combinatorial Search

Invited Speakers

We are pleased to have the following invited speakers at SoCS 2017:

Daniel Borrajo
Talk Title: Diverse, Adaptive and Declarative Problem Solving


This talk will cover three distinct concepts that have partly driven problem solving within AI. Diversity relates to the variety of problem solving tasks, and the known fact that there is no single tool or technique that works best for all tasks. Recently, Diversity has been addressed by portfolios, but it has also been considered by many other works as in bidirectional search, or combinations of heuristics. Given that we need to solve tasks with diverse properties and structure, many approaches have used Adaptation. Again, it has taken many different forms from machine learning to automated parameterization. Finally, Declarative representations are a key principle of (part of) AI research that are based on separating the definition of domain and problem models (what) from the problem solvers (how). I will argue that the joint consideration of these three concepts can lead to flexible and high performance problem solving techniques.

About the Speaker:

Daniel Borrajo received his PhD in 1990 and is currently a full professor at Universidad Carlos III de Madrid, Spain. His research interests lie on automated planning, covering topics from machine learning applied to problem solving (learning action models, portfolios or heuristics) to more recent work on multi-agent planning, or goals management. He has been a member of the SoCS and ICAPS councils.

Shin-ichi Minato
Talk Title: Power of Enumeration -- BDD/ZDD-Based Techniques for Solving Combinatorial Problems


BDD (Binary Decision Diagram) is a classical data structure for representing a Boolean function. BDD-based algorithms were developed mainly for VLSI logic design in early 1990s. ZDD (Zero-suppressed BDD) is a variant of BDD, customized for representing a set of combinations, often appear in solving combinatorial problems. BDDs and ZDDs have become more widely known since D. Knuth intensively discussed them in his famous series of books in 2009. Although a quarter of a century has passed since the original idea of using BDDs by R. Bryant, there are still many interesting research topics related to BDDs/ZDDs. One of the most important topics is a fast algorithm of constructing a ZDD which enumerates all the paths in a given graph structure. This work is important because many kinds of practical problems are efficiently solved by some variations of this algorithm. In this talk, I will give an overview of the brief history and the basic techniques on BDDs/ZDDs. We then look over some recent research topics to show "the power of enumeration" for solving real-life combinatorial problems.

About the Speaker:

Shin-ichi Minato is a Professor at Graduate School of Information Science and Technology, Hokkaido University, Japan. His research interest includes efficient data-structures and algorithms for manipulating very large-scale data of discrete structures. He received the B.E., M.E., and D.E. degrees from Kyoto University in 1988, 1990, and 1995, respectively. He had been working at NTT Research Laboratories since 1990 until 2004. He was a Visiting Scholar at Stanford University in 1997. He joined Hokkaido University in 2004 and has been a Professor since 2010. He served the Research Director of "ERATO," a nation-wide research project of Japan, from 2009 to 2016. He published "Binary Decision Diagrams and Applications for VLSI CAD" (Kluwer, 1995).