Showing papers by "Fabio Crestani published in 2000"

User's perception of relevance of spoken documents

Abstract: We present the results of a study of user's perception of relevance of documents. The aim is to study experimentally how users' perception varies depending on the form that retrieved documents are presented. Documents retrieved in response to a query are presented to users in a variety of ways, from full text to a machine spoken query-biased automatically-generated summary, and the difference in users' perception of relevance is studied. The experimental results suggest that the effectiveness of advanced multimedia Information Retrieval applications may be affected by the low level of users' perception of relevance of retrieved documents.

Searching the Web by constrained spreading activation

Abstract: Intelligent Information Retrieval is concerned with the application of intelligent techniques, like for example semantic networks, neural networks and inference nets to Information Retrieval. This field of research has seen a number of applications of Constrained Spreading Activation (CSA) techniques on domain knowledge networks. However, there has never been any application of these techniques to the World Wide Web. The Web is a very important information resource, but users find that looking for a relevant piece of information in the Web can be like ‘looking for a needle in a haystack’. We were therefore motivated to design and develop a prototype system, WebSCSA (Web Search by CSA), that applied a CSA technique to retrieve information from the Web using an ostensive approach to querying similar to query-by-example. In this paper we describe the system and its underlying model. Furthermore, we report on an experiment carried out with human subjects to evaluate the effectiveness of WebSCSA. We tested whether WebSCSA improves retrieval of relevant information on top of Web search engines results and how well WebSCSA serves as an agent browser for the user. The results of the experiments are promising, and show that there is much potential for further research on the use of CSA techniques to search the Web.

Soft computing in information retrieval: techniques and applications

Abstract: Information retrieval (IR) aims at defining systems able to provide a fast and effective content-based access to a large amount of stored information. The aim of an IR system is to estimate the relevance of documents to users' information needs, expressed by means of a query. This is a very difficult and complex task, since it is pervaded with imprecision and uncertainty. Most of the existing IR systems offer a very simple model of IR, which privileges efficiency at the expense of effectiveness. A promising direction to increase the effectiveness of IR is to model the concept of "partially intrinsic" in the IR process and to make the systems adaptive, i.e. able to "learn" the user's concept of relevance. To this aim, the application of soft computing techniques can be of help to obtain greater flexibility in IR systems.

Exploiting the Similarity of Non-Matching Terms at RetrievalTime

Abstract: In classic Information Retrieval systems a relevant document will not be retrieved in response to a query if the document and query representations do not share at least one term. This problem, known as “term mismatch”, has been recognised for a long time by the Information Retrieval community and a number of possible solutions have been proposed. Here I present a preliminary investigation into a new class of retrieval models that attempt to solve the term mismatch problem by exploiting complete or partial knowledge of term similarity in the term space. The use of term similarity enables to enhance classic retrieval models by taking into account non-matching terms. The theoretical advantages and drawbacks of these models are presented and compared with other models tackling the same problem. A preliminary experimental investigation into the performance gain achieved by exploiting term similarity with the proposed models is presented and discussed.

Effects of word recognition errors in spoken query processing

Abstract: The effects of word recognition errors (WRE) in spoken document retrieval have been well studied and well reported in recent information retrieval (IR) literature. Much less experimental work has been devoted to studying the effects of WRE in spoken query processing in IR. It is easy to hypothesize that given the typical length of the user query, the effects of WRE in spoken queries on the performance of IR systems must be destructive. The experimental work reported tests this. The paper reports on the background of the study, on the construction of a suitable test collection, on the first experimental results obtained and on the limitations of the study. The results show that classical IR techniques are quite robust to considerably high levels of WRE rates in spoken queries (roughly below 40%), in particular for long queries.

The Visual Book and the Hyper-TextBook: two electronic books one lesson?

Abstract: This paper presents the results of two separate studies into electronic book production The Visual Book (Landoni, 1997) which explored the importance of the visual component of the book metaphor in the production of "good" electronic books, and the Hyper-TextBook (Crestani and Melucci, 1998a) which instead concentrated on the importance of hypertext models for the automatic production of functional electronic books Both studies started from similar considerations on what kinds of paper books are more suitable for being translated into electronic form and both identified as target publications which are meant to be used as references more than sequentially read by users The result of these two research strands is critically presented in this paper, with the aim of integrating them in a more general research work aimed at providing a complete methodology on how to produce electronic books in a cheap and fast way, so that they are visually appealing and at the same provide the user with a set of desirable searching and browsing functionalities

Imaging and Information Retrieval: Variations on a Theme∗

Abstract: (Standard) Imaging is a method for the revision of probability functions originally proposed in the philosophy of language as a semantics for conditional logic. Recently, Standard Imaging and a variant of it, called General Imaging, have successfully been applied to the estimation of the probability of relevance in Information Retrieval (IR) by Crestani and van Rijsbergen. The experimental results they have obtained show that these methods perform better than a number of more established approaches, such as retrieval by joint or conditional probability. In this paper we report preliminary results on three main generalisations of these methods and their application to IR. These generalisations are orthogonal (and they may thus be freely combined), as they address three orthogonal issues in the probability kinematics of Imaging. The first generalisation, that we call Proportional Imaging, is a variation of General Imaging that is better suited to those cases in which similarity between “possible worlds” has a quantitative nature; this is indeed the case in the application of Imaging methods to IR, where keywords play the role of possible worlds. The idea that underlies Proportional Imaging is that the probability of an A-world w should be distributed to all A-worlds wi in a way that is proportional to the degree of similarity between w and the wi’s. ∗This work has been carried out in the context of the project FERMI 8134 “Formalisation and Experimentation in the Retrieval of Multimedia Information”, funded by the European Community under the ESPRIT Basic Research scheme.

Probabilistic Learning by Uncertainty Sampling with Non-Binary Relevance

Abstract: We present a learning model for probabilistic learning in information retrieval and information filtering which is based on the concept of “uncertainty sampling”. Uncertainty sampling is a technique that exploits user relevance feed-back both for relevant and non-relevant documents. In particular, relevance sampling uses those documents whose relevance is most uncertain to speed up the learning of the user relevance criteria. We extend the use of uncertainty sampling by considering multiple levels of relevance and we show how this new learning model for information retrieval and filtering could be evaluated using collections with non-binary relevance assessments.

Proceedings of the Third European Summer-School on Lectures on Information Retrieval-Revised Lectures

Abstract: From the Publisher: "This book presents 12 revised lectures given at the Third European Summer School in Information Retrieval, ESSIR 2000, held at the Villa Monastero, Varenna, Italy, in September 2000. This first part of the book is devoted to the foundations of IR and related areas; the second part on advanced topics addresses various current issues, from usability aspects to Web searching and browsing." "This book is written for researchers starting out in IR, for scientists and engineers in corporate R&D who wish to learn more about this increasingly important area, and for people working on topics related to the management of information on the Internet."--BOOK JACKET.

Spoken Query Processing for Information Access in Digital Libraries

Abstract: We briefly outline the ongoing research at Strathclyde University on the use of spoken query processing for information access in digital libraries.

Logical and uncertainty models for information access: current trends

Abstract: The current trends of research in information access as emerged from the 1999 Workshop on Logical and Uncertainty Models for Information Systems (LUMIS'99) are briefly reviewed in this paper. We believe that some of these issues will be central to future research on theory and applications of logical and uncertainty models for information access.