WebMedia 2008 || Vila Velha-ES

As seguintes palestras de convidados nacionais e internacionais estão programadas para os eventos WebMedia, SBSC, WBPM, TIL e LA-WEB:

Experience in World Wide Web (Web) accessibility has taught us: to think about small bespoke solutions; to tailor interaction and requirements to the user and job at hand; to value high data interoperability; to realise that large enterprise systems, become unmanageable and unable to change at the speed required by both users and technology; and finally, to value heterogeneity. Indeed, with the advent of Workflows, RPCs, RESTful services, and Cloud Computing we can also see these viewpoints becoming more common in mainstream thought. Here, we use Bernstein's concept of 'NeoVictorian Computing' as a counterfoil to Andriole's new viewpoint of 21st century software development. We extrapolate from the Web development model into corporate and enterprise systems and propose a architecture of client based heterogeneous applications each tailored to a specific user, and their job, with highly interoperable data, controlled by workflow's that are transferred with the data itself. We discount the new client-computing fad, as this really means centrally controlled, sometimes unavailable, old style enterprise systems. We suggest that by moving toward user centred agile systems we follow the conceptual, if not the technological, underpinnings of the Web. In this case we realise that Web developers are in a privileged position to shape and push forward this new kind of software architecture and the 'craft' based approaches which will drive it.

Abstract
Conceptual Modeling is a discipline of great importance to several areas in Computer Science such as Domain and Knowledge Engineering, Organizational Modeling, Information Systems Design, Database Design, Knowledge Management, among many others. In particular, what is termed a Domain Ontology in Computer Science is a special type of Conceptual Model. In recent years, there has been a growing interest in the development and use of domain ontologies, strongly motivated by the Semantic Web initiative. However, as we argue in this talk, an approach for ontology engineering uniquely based on the modeling languages adopted in the Semantic Web (e.g., OWL, RDF) is insufficient to address a number of representation and semantic interoperability problems that arise in open and dynamic scenarios (such as, for instance, the Semantic Web itself). In this talk, we argue that these languages should be complemented by a language and methodology based on a Foundational Ontology, i.e., a domain-independent common-sense theory constructed by aggregating suitable contributions from areas such as philosophical ontology and logics, cognitive science and linguistics. Moreover, we discuss the requirements for a theoretically well-founded modeling language which are needed to meet the desiderata for a general conceptual modeling and ontological representation language. Finally, we briefly present advanced conceptual modeling techniques (such as design patterns and methodological guidelines) based on principled ontological foundations and show how they can be used to solve some classical and recurrent conceptual modeling problems that (re)appear in concrete application scenarios.

Palestrante: Fabio Casati, departamento de Ciência da Computação, Universidade de Trento (Itália)
Título da Palestra: Resource Lifecycle Management: BPM at work in the social web.

Abstract
The social web is (also) about a web of resources that are collaboratively managed, often in a fairly unstructured fashion. Examples are wikis or google docs. The same philosophy is spreading to "business resources", such as project schedules and project deliverables, often shared and edited collaboratively, though often with a little more structure. In open source software development, in a way a precursor to Web 2.0, the "software resources" are managed similarly. Some aspects common to all these resources is that 1) they go through some lifecycle (which may or may not be structured and defined); 2) even when defined, the lifecycle varies frequently, and this variation is not an "exception", is normal business life (just imagine defining a strict lifecycle for your project deliverables and sticking to it 100%....); 3) humans are in control: managing resources is not about automating the execution of their lifecycle. I certainly would not want a BPM to manage my activity of writing a paper. However, it is about being able to initiate actions when needed (submission for review, testing of software) and to monitor the status and history of a possibly large set of resources that I am managing (e.g., a set of deliverables for my project).
In this talk I will present a resource lifecycle management system that can manage anything that can be referred to by a URI, that can be used by any skilled web user even without programming skills. It's key aspects are the simplicity of the model, the embracing of the inherent "unstructuredness" of the life of (Web) resources, the ability to monitor status and history of all managed resources, the ability to make changes easily on the fly, and the balance between human and automated control. In particular, in the RLM system, the "workflow engine" are in fact the humans managing the resources, initiating automated actions if and when needed. The approach is inspired in part by BPM systems and in part by the philosophy of the Web.The RLM system will be hosted and available for use by anybody who wants to model the lifecycle of any resource, to automate certain (resource-specific) actions (e.g., submission for review), and to observe the evolution of a set of resources of interest. The applications are countless, but examples include project deliverables (which typically vaguely follow a lifecycle based on a project quality plan), collaborative editing on wikis, writing of code and documentation, or composing, arranging, and recording a song.
.

Palestrante: Ruy Luiz Milidiu, Departamento de Informática, Pontifícia Universidade Católica RJ (PUC-RJ)
Título da Palestra: F-EXT: a framework for semantic facts extraction.

Abstract
The application of Machine-learning techniques to fact extraction has been reported since the formalization of the Template Element Task at the Message Understanding Conference (MUC-6, MUC-7) and Automatic Content Extraction meetings (NIST, 2000). This talk reports on several applications of the F-EXT Framework. F-EXT provides the basic architecture and functionality for the design and implementation of machine learning based Fact Extractors. We consider Fact extraction a variety of semantic extraction applications like named relation extraction and semantic role labelling. F-EXT is also used as a basic machine learning tool for text preprocessing tasks, that are key to the construction of more advanced tasks like accurate relation extractors. The main component of F-EXT relies on its core machine learning engines which are Hidden Markov Models, Support Vector Machines, Decision Trees, Transformation Based Learning and Entropy Guided Transformation Learning. We report on the application of F-EXT to develop several NLP preprocessors: Portuguese POS tagger, English noun phrase chunker, English text chunker, Portuguese noun phrase chunker, Hindi text chunker and Portuguese named entity recognizer. We also report on the application of F-EXT to design and implement the following four relation extractors: Person-Affiliation, Person-Quotation, Portuguese Person-Appositive and predicate-arguments. F-EXT offers a straightforward usability that enables low effort fact extractors construction. Its advanced machine learning algorithms allow the construction of state-of-the-art competitive natural language processing tools.

Palestrante: Clarence Ellis, Collaboration Technology Research Group, Universidade do Colorado (EUA)
Título da Palestra: Election2008-USA: A Massively Multiplayer Game

Abstract
This talk will begin with a bit of social commentary on the current ongoing US presidential election process. This will lead into presentation of a proposed new election process called Humane Rational Democracy that can be implemented in a staged fashion through the use of social networking and massively multiplayer Internet games. We argue that the time and technology are right for this experiment. An overview of the process, the games, and the selection process for electors will be presented, with time allocated to discussion, questions, and opinions.

Palestrante: Fabrizio Silvestri, CNR (Itália)
Título da Palestra: Past Searches Teach Everything: Including the Future!

Abstract
"History Teaches Everything, Including the Future" wrote Alphonse de Lamartine in the nineteen century. Even if history cannot be really considered a predictive science, historical information can successfully be used in many fields. In this talk I will discuss Web Search Engines and their Query logs, which contain historical information about past usage of such systems. Query Logs are nowadays considered as one of the most important sources of knowledge that is routinely exploited to improve performance in terms of both effectiveness and efficiency of Web Search Engines. The techniques I am going to review are mainly focused on enhancing efficiency of real-world distributed search systems. I will present some of the most interesting results obtained in the last five years by us at the High Performance Computing Laboratory in Pisa in collaboration with research labs Worldwide. I will start by showing state-of-the-art caching techniques for search results and posting lists and ways to optimally combine them. Next, I will show applications of data mining techniques aimed to speed-up the processing of queries in both document-based, and term-based distributed search engines. I will conclude by presenting two new results. One is related to caching in similarity-based multimedia search systems. The latter is about caching for efficient snippet generation on a distributed repository of documents.

Palestrante: Mariano Consens, Information Engineering, MIE and CS, Universidade de Toronto (Canadá)
Título da Palestra: Managing Linked Data on the Web: the LinkedMDB Showcase

Abstract
The Linking Open Data community project is extending the Web by encouraging the creation of interlinks (RDF links between data items from different datasets identified using dereferenceable URIs). This emerging web of linked data is closely intertwined with the existing web, since structured data items can be embedded into web documents (e.g., RDFa or microformat encoded data), and RDF links can reference classic web pages. Abundant linked data justifies extending the capabilities of web browsers and search engines, and enables new usage scenarios, novel applications and sophisticated mashups.

This promising direction for publishing data on the web brings forward a number of challenges. While existing data management techniques can be leveraged to address the challenges, there are unique aspects to managing web scale interlinking. In this presentation, we describe two specific challenges; achieving and managing dense interlinking, and describing data, metadata, and interlinking within and among linked web datasets. Approaches to solve these challenges are showcased in the context of LinkedMDB, the first open linked dataset for movies. LinkedMDB has a large number of interlinks (over a quarter of a million) to other open linked datasets, as well as RDF links to movie-related webpages.

Abstract
Today, with the proliferation of affordable computing, people use multiple devices to fulfill their information needs. However, designers approach each device platform individually, without accounting for the other devices that users may also use. In many cases, the applications on all the user's devices are designed to be functional replicates of each other, often with an emphasis on keeping their form and function consistent with one another. We argue that this emphasis on recreating consistent clones on each platform should not be the dominant concern for designers. In this presentation, we present the idea of a personal information ecosystem, an analogy to biological ecosystems, which allows us to discuss the inter-relationships among users devices to fulfill their information needs. Using the examples of now ubiquitous web technologies on different platforms, we discuss how considering the user's ecosystem of devices as a whole as a design target lends to better user experience and encourages designers to tackle the more important concern of seamless task migration across devices.

Abstract
- Conceitos básicos que regem a colaboração na Internet
- Governança histórica dos recursos centrais, IAB, IANA, IESG, IETF
- Mudanças no final do anos 90: ICANN
- Cúpula de Túnis, WSIS, WGIG e IGF (Atenas e Rio de Janeiro)
- Nomes de Domínio. ccTLDs, gTLDs, IDN
- A transição para inclusão do IPv6
- O CGI, o registro brasileiro, DNSSEC

Palestrante: : Ricardo Baeza-Yates, VP of Yahoo! Research
Título do Tutorial: Introduction to Web Mining

Abstract
The Web continues to grow and evolve very fast, changing our daily lives. This activity represents the collaborative work of the millions of institutions and people that contribute content to the Web as well as the one billion people that use it. In this ocean of hyperlinked data there is explicit and implicit information and knowledge. Web Mining is the task of analyzing this data and extracting information and knowledge for many different purposes. The data comes in three main flavors: content (text, images, etc.), structure (hyperlinks) and usage (navigation, queries, etc.), implying different techniques such as text, graph or sequence mining. Each case reflects the wisdom of some group of people that can be used to make the Web better. For example, user generated tags in Web 2.0 sites. In this talk we walk through the mining process and show several applications, ranging from Web site design to search engines.