ESSIR'03
Lectures


Welcome
PHOTOS!!!
Programme
Lecturers
Lectures
email webmaster

 Introduction to Information Retrieval

Lecturer: K. J. van Rijsbergen

PDF slides

Description of lesson.


 Formal Models of Information Retrieval

Lecturer: D. Hiemstra

PDF slides

A Tutorial to Formal Models of Information Retrieval.
In this lecture, I will give a "grant overview" of formal models of information retrieval with special attention to probabilistic approaches. After the course, students will have enough information to choose the formal information retrieval framework that serves their research the best: They will know the strengths and weeknesses of the most influencial retrieval models, and they will have some sense of the "history of information retrieval modelling". An important part of this lecture will cover the recently developed language modelling approach to information retrieval. I will show several of these language modelling approaches, including models for cross-language information retrieval, adaptive filtering and web search.


 Logical Models of Information Retrieval

Lecturer: Fabio Crestani

PDF slides

The use of different logics in Information Retrieval enables to formulate models that are more general than the classical IR models. In fact, logical models of IR not only enable to represent in a more previse way the document informative content and the user information need, but they also enable the representaion of some advanced features of IR systems, such as hypermedia links, multimedia content, and users knowledge. Logics also provides a common approach to the integration of IR systems with logical database systems. Finally, logic makes it possible to reason about an IR model and its properties.

This lecture will give an overview of how different logical systems can be used for the formulation of more advanced models for the representation and retrieval of information.


 Web Retrieval

Lecturers: M. Agosti and M. Melucci

PDF slides of M. Agosti
PDF slides of M. Melucci

The talk is going to address different aspects of the retrieval of information from Web pages and documents. The Web is formed by static pages and dynamic ones, which give access to many different types of databases and digital libraries. When the retrieval of information is performed on static pages, many are the relationships with traditional information retrieval. When the retrieval of information is obtained using tools which access engines which extract information from databases and digital libraries different types of selection of information is conducted. The talk addresses both types of situations.

One of the main differences between conventional document collections and the Web is the availability of links. Links are an important source of evidence to design and implement methods for the retrieval of information overcoming the drawbacks of text retrieval, e.g. polysemy and synonymy. Thanks to links, the retrieved Web pages can be ranked using the topology of the graph which they belongs to estimate a sort of "importance". The talk illustrates the most known methods.


 Users: the Central Context in Interactive Information Retrieval

Lecture: P. Ingwersen

PDF slides

Users as actors in interactive information retrieval (IIR) are seen in the contexts of their perceived work tasks and information seeking behaviour, and as contexts to the IR system. The paper models IIR processes by demonstrating a variety of approaches, ranging from Ingwersen´s cognitive communication model for IR interaction, over Saraceveic´ stratified model which includes a typology of relevance conceptions, to Borlund´s model of work task perception, information need development and relevance assessments. Other associated models and perspectives of IIR are discussed when appropriate to the major focus points of the contribution: work/search task - interest conceptions; information need development and typology; understanding of relevance in IIR; and experimental problems in IIR.


 Information Retrieval and Structured Documents

Lecturer: Y. Chiaramella

PDF slides

Description of lesson.


 Evaluation in Information Retrieval (including TREC)

Lecturer: S. Robertson

PDF slides

The field now known as "information retrieval" long predates computer-based systems, and there has been a tradition of evaluating information retrieval systems for over 40 years (again predating computer-based system). In part, this tradition is embedded in an empirical, laboratory-based paradigm, in which evaluation is a process of measurement, requiring the definition of quantitative measures of performance. This tradition is now best represented by the TREC programme. The primary basis for quantitative measurement is normally a set of judgements of the "relevance" of documents to information needs. This in itself is of course a qualitative judgement; the quantitativeness arises from counting documents, in categories defined by relevance and other properties.

At the same time, a diversification of this tradition has taken place, and attempts are made to observe information retrieval activities in real-life environments, perhaps in a more qualitative fashion. There is some tension between the laboratory and real-life evaluation ideas, but they have complementary functions. Some real-life studies go beyond the immediate observation of acts of retrieval, and consider (say) the work context of this task. There are also attempts to introduce alternative task-context models back into laboratory experiments.

The object of this session will be to give participants an overview of the principles of evaluation in IR, an appreciation of the difficulties of conducting experiments, and some of the results of past studies.


 Multilingual Information Retrieval

Lecturer: J.-Y. Nie

PDF slides

Description of lesson.


 Perception, Signal, Images

Lecturers: A. Guérin-Dugué and C. Berrut

PDF slides

Description of lesson.


 User Needs in Video

Lecturer: B. Bachimont

PDF slides

Audiovisual documents are more and more in a digital format, be they digital born or obtained from a digitisation process. This simple fact has two main consequences : audiovisual documents are used on the same medium as the metadata that describe them ; they also can be used in the context of other digital documents that have different semiotic formats (texts, sounds, graphics, etc.). The lesson will concentrate on the implications of this revolution from the application point of view on one hand and the needed technologies on the other. We will show how metadata and indexing technics are very promising in the digital era while being still hard to use in real life applications.


 Models, Tools for Videos

Lecturer: G. Quénot

PDF slides

Description of lesson.


 Video Searching and Browsing

Lecturer: A. Smeaton

PDF slides

As computers have become more and more powerful, and as available storage has become cheaper and the volumes increased, and as computer networks have become faster and faster, we have seen the growth of more and more information in digital format. In recent years, this growth has included the growth in the use of digital formats for capturing, storing, and transmitting video information. Digital format for video information can now be captured easily and transmitted and rendered on desktop and handheld mobile devices. In this lecture we address the current state of the field in the area of content-based operations to operate on digital video information. As background information, and to give some context for the difficulties associated with content access to video, we include some materials on video compression algorithms and approaches and we then provide details on feature extraction, indexing, searching, browsing, summarisation and automatic linking in digital video information. Our examples will be taken from the TREC track on video IR (known as TRECVID) and from our own work on the Físchlár system.



This site is optimized for MS Internet Explorer 5.5 and higher, Netscape 6.0 and higher, Opera 6.0 and higher. It uses cascading style sheets, javascripts and java applets.
Date of last modification: 02/02/2004
email webmaster