Francesco Scarcello is a full professor of computer science (SSD IINF-05/A) at the University of Calabria, where he currently serves as Deputy Rector and Delegate for Teaching and Learning, too.
He is an EurAI Fellow and an AAIA Fellow.
Short Bio, and main experiences.
Francesco Scarcello received the PhD degree in Computer Science from the University of Calabria in 1997; he is with University of Calabria since 1999. Before that, he was recipient of two grants from the Italian National Research Council (CNR). Also, during those years, he has been visiting the Department of Information Systems at the Vienna University of Technology.
After his first position as an assistant professor, he got the position of associated professor in 2001, and the current position of full professor since 2013.
Among his various roles at University of Calabria, let us recall that he is Delegate of the Rector for Teaching and Learning since 2014, and also Vice Rector since 2020.
He is a member of the Management Board of the CINI Laboratory of Artificial Intelligence and Intelligent Systems, and he is a member of the PhD Committee of the National PhD in Artificial Intelligence.
Professor Scarcello serves on program committees and as a reviewer for many international conferences and journals. Until 2020, he served as an Associated Editor of the Artificial Intelligence journal (Elsevier).
Main Research Interests.
The main interests of Francesco Scarcello are in computational complexity, graph and hypergraph theory, constraint satisfaction, game theory, logic programming, knowledge representation, non-monotonic reasoning, and database theory. He has extensively published in all these areas in leading conferences and journals.
In database theory, he solved some open problems on the complexity of acyclic conjunctive queries (Journal of the ACM 48(3):431-498,2001), and on some techniques to identify tractable classes of queries. From these studies the idea of a new notion of degree of cyclicity for hypergraphs, called hypertree-width came out. It turned out that this notion is the best technique for identifying tractable classes of hard problems, in that it is able to single out the largest islands of tractability among all other methods previously known in the literature (Artif. Intell. 124(2): 243-282, 2000, J. Comput. Syst. Sci. 64(3): 579-627, 2002, J. Comput. Syst. Sci. 66(4): 775-808, 2003, J. Comput. Syst. Sci. 73(3): 475-506, 2007). Hypertree decompositions, being a general (hyper)graph notion, find many applications in various problems from different fields, in particular in Artificial Intelligence (Inf. Comput. 252: 201-220, 2017; SIAM J. Comput. 46(3): 1111-1145,2017).
The notions and the algorithms proposed in his Phd thesis for disjunctive logic programing have been used for developing DLV, a widely used knowledge-base management system based on Answer Set Programming (Inf. Comput. 135(2): 69-112, 1997, ACM Trans. Comput. Log. 7(3): 499-562, 2006). A spin-off company, DLVsystem srl, was founded in 2005 to continue the development and improvement of DLV. Its current implementation is able to deal with Big Data and to interact with other paradigms of Artificial Intelligence (Machine Learning, Deep Learning, Computer Vision).
Francesco Scarcello is also interested in multi-agents and game theory. He first studied the complexity of computing Nash Equilibria in Strategic Games (JAIR, 24:357-406, 2005; TCS (38-40):3901–3924, 2009), for which he has shown that, even in restrictive settings, determining whether a game has a pure Nash equilibrium is NP-complete (and ΣP2-complete for strong equilibria). Tractable cases are also identified, by considering quantitative and qualitative restrictions of the way each player’s move depends on moves of other players.
He then studied cooperative games, in particular coalitional games, where each set of agents (coalition) is associated with a potential worth, and the problem is how to assign values to each agent in a way that is fair (and possibly meets further requirements). He studied the main solution concepts for these games in a series of publications, where he also solves some open problems about the precise complexity of these notions (Artif. Intell. 278, 2020; ACM Trans. Comput. Theory 7(1): 3:1-3:52, 2014; Artif. Intell. 175(12-13): 1877-1910, 2011). He also worked on applications of these game-theoretic notions to wireless communications and sharing services (IEEE Trans. Mob. Comput. 13(10): 2185-2198, 2014; IEEE Trans. Veh. Technol. 63(1): 478-484, 2014; Comput. Networks 57(9): 1955-1973, 2013; IEEE Trans. Wirel. Commun. 10(8): 2566-2576, 2011).
A related line of research is about mechanism design, in particular allocation problems, and coalitional games induced by matching problems (JAIR 49: 403-449, 2014; Artif. Intell. 278, 2020, AAAI 2020: 2006-2013).
With this respect, algorithms for the computation of the Shapley value have been designed and implemented for an interesting game-theoretic application to the Italian VQR research assessment exercise, in particular for assigning to each researcher a value that is “fair” (in a precise sense), with respect to her/his actual contribution the university score (J. Informetrics 13: 87-104, 2019; J. Exp. Theor. Artif. Intell. 30: 505-524, 2018).
Projects and Technology Transfer
Francesco Scarcello is also active in technology transfer and research projects, he is a founder of three university spin-off companies, namely DLVsystem srl, Artémat srl and EVO-BI srl, and participated in research projects dealing with database, artificial intelligence, and knowledge representation systems.
Let us briefly recall some projects where he played the role of scientific coordinator.
- Industrial Project on Business Games, with Artémat srl: research on the development of Business Games specifically designed for knowledge-based economies, in particular for the ICT market. The research focused on a framework for the knowledge representation of market models, and on suitable ways of playing on limited devices, such as the mobile phones.
- Data Integration for Business Intelligence, with the Competence Center ICT-SUD, industrial partners, and health experts from the Azienda Sanitaria Provinciale of Cosenza (ASP-CS): research on the integration of data from different sources, possibly incomplete and inconsistent, and on techniques for doing Business Intelligence with such a data. The challenging use case was the development of the Cancer Registry for the ASP-CS.
- Talent Hunter Technology, with DLVsystem srl and Artémat srl: research on knowledge representation and automatic extraction of semantic data from curricula vitae et studiorum, with an application to a Human Resources (HR) tool to identify the most promising candidate profiles.
- Open Source Monetization Ecosystem (OSME), with HicTech srl, Artémat srl, and Talent Garden: design techniques, models, and suitable licenses to develop a digital ecosystem for the Open Source Software, which aims at supporting open-source developers by recognizing and rewarding their actual contribution to projects. The research faced both economical modeling and technical issues related to software evaluation and transparent deploying on cloud platforms.
Impact of Research and Awards.
Francesco Scarcello has extensively published in leading conferences (FOCS, PODS, IJCAI, etc.) and journals (Journal of ACM, Artificial Intelligence, SIAM Journal on Computing, Journal of Computer and System Sciences, Information and Computation, IEEE Transactions on Knowledge and Data Engineering, etc.). He gave invited talks and tutorials in international conferences, such as the International Conference on Principles of Knowledge Representation and Reasoning and the International Joint Conference on Artificial Intelligence.
According to Google Scholar, his publications received more than 5700 citations; his H-index is 34, and his i10-index is 51.
He received the Best Paper Award for his work on Fixed-Parameter Complexity in AI and Nonmonotonic Reasoning, at the 5th International Conference on Logic Programming and Nonmonotonic Reasoning (LPNMR ’99).
His work on the complexity of Nash equilibria in non-cooperative games received in 2008 the IJCAI-JAIR Best Paper Prize, which is awarded to an outstanding paper published in Journal of Artificial Intelligence Research in the preceding five calendar years.
His paper ``Hypertree Decompositions and Tractable Queries'' received in 2009 the PODS Alberto O. Mendelzon Test-of-Time Award, which is awarded every year to a paper, or a small number of papers published in the proceedings of the Symposium on Principles of Database Systems (PODS) ten years prior that had the most impact in terms of research, methodology, or transfer to practice over the intervening decade (PODS is the premier international conference on the theoretical aspects of database systems).
In fact, this work, together with a number of subsequent related publications, had a significant impact both in Artificial Intelligence and Database Theory, and many other researchers worked on this subject, producing important results, both in theory and in practical applications. Hypertree Decompositions are described both in books for researchers, such as Constraint Processing (by Rina Dechter, Morgan Kaufmann ed., 2003) and in general books used for teaching Artificial Intelligence in most universities, such as Artificial Intelligence – A Modern Approach (Stuart Russell - Peter Norvig, Pearson ed., 2010).
Another important reward for his work is the invitation, in 2016, to contribute with the paper “Hypertree Decompositions: Questions and Answers” to the very first edition of the PODS special session “Gems of PODS”, which aims at promoting the knowledge of the most significant works in database theory.
