Seminars at the Faculty of Informatics

The Faculty of Informatics is pleased to announce a seminar given by Alessandro Bozzon

DATE: Friday, October 5th, 2012
PLACE: USI Università della Svizzera italiana, room A33, Red building  (Via G. Buffi 13)
TIME: 09:30

ABSTRACT:
The retrieval and management of Web data is becoming a more and more complex problem, due to the amount of available contents, to the diversity of the information sources, and to the evolving expectations of users. In this talk I will address some of the current research opportunities and challenges in Web data management, and I will present my research agenda, which focuses on the following research questions:         

+ *Given the massive amount of information that are daily created on the Web, how to guarantee scalability of data processing and analysis? How to extract value and knowledge from the incredible amount of information that is produced, in a reasonable time?* "Order-aware Reasoning and Data Management" (ORDRING) is a recent proposal for efficient data processing that advocates the exploitation of orders in the data (or in the query) to narrow the solution space and, ultimately, speed-up performance. In my talk I will illustrate the rationale behind the ORDRING vision, and I will present SPARQL-RANK, an extension of the standard SPARQL algebra and execution model that, by considering ORDER as a first-class property of data and queries, yields significant performance boost for order-aware SPARQL queries.

+ *Are state-of-the-art information seeking systems able to support users in their search tasks? How Web data sources can be queried so to support long-lived search processes?* An important trend in information retrieval is the support for search processes that go beyond one-time interactions, while allowing Web users to progressively seek for information from several sources. Such an information seeking behavior, typically referred to as "Exploratory Search", is not supported by traditional Web search engines. I will present a general-purpose search framework that formalizes exploratory search behaviors while supporting search processes that span over multiple, searchable data sources. Experiments with users confirmed the need for a new class of exploratory search systems, and shown the utility and effectiveness of our approach.

+ *How Web users can be (actively or passively) engaged in the curation, verification, analysis, and enrichment of contents?* Human Computation is a recently revamped discipline that advocates the usage of human processing to support and/or substitute computers in solving problems that they cannot solve yet. In my talk, I will discuss how Web users can be engaged in human computation activities; I will describe an high-level, conceptual framework, which takes in input a problem solving scenario, formalized as program executable cooperatively by both humans and machines, and I will introduce "CrowdSearcher", a system architecture bridging conventional search experiences to crowdsearching and social network exploitation.

 

BIO:
Alessandro Bozzon is a Post-doctoral researcher in the "Web and Database group" at  Politecnico di Milano, Italy. He obtained his PhD in Computer Science from Politecnico di Milano in 2009, with a dissertation on model driven approaches for the design, development and automatic code generation of Search Based Applications. His current research interests are on data- and information management aspects, with specific focuses on databases; data integration; the Semantic Web; human computation; data interaction, visualization, and diversification; and textual and multimedia information retrieval. Alessandro has been actively involved in the definition and execution of several Italian and European research projects in the above areas, where he served as responsible and coordinator for several research, implementation and demonstration work-packages.

HOST: Prof. Mauro Pezzè