In recent years, interface agents have been developed to assist users with various tasks. Some systems employ machine learning techniques to allow the agent to adapt to the user's changing requirements. With the increase in the volume of data on the Internet, agents have emerged which are able to monitor and learn from their users to identify topics of interest. One such agent, described here, has been developed to filter mail messages. We examine the issues involved in constructing an autonomous interface agent which employs a learning component, and explore the use of two different learning techniques in this context. Submitted to Applied Artificial Intelligence Journal. October 26, 1 INTRODUCTION 1 1 Introduction Agents were once seen as anthropomorphic entities which would assist users with daily tasks. They could be used, for example, to locate information of interest to their user (Kay 1984). Ten years later, many definitions of agents have been proposed. The basic concept of ...

In this paper we investigate the applicability of on-line learning algorithms to the real-world problem of web search. Consider that web documents are indexed using Boolean features. We first present a practically efficient on-line learning algorithm TW2 to search for web documents represented relevant features. We then design and implement WebSail, a real-time adaptive web search learner, with TW2 as its learning component. WebSail learns from the user's relevance feedback in real-time and helps the user to search for the desired web documents. The architecture and performance of WebSail are also discussed. 1.

Rocchio's similarity-based Relevance feedback algorithm, one of the most important query reformation methods in information retrieval, is essentially an adaptive supervised learning algorithm from examples.

Information retrieval systems provide access to collections of thousands, or millions, of documents, from which, by providing an appropriate description, users can recover any one. Typically, users iteratively refine the descriptions they provide to satisfy their needs, and retrieval systems can utilize user feedback on selected documents to indicate the accuracy of

We consider the problem of selecting the articles of experts' interest from a literature database with the assistance of a machine learning system. For this purpose, we propose the rough reading strategy which combines the experts' knowledge with the machine learning system. For the articles converted through the rough reading strategy,we employ the learning system BONSAI and apply it for discovering rules whichmay reduce the work of experts in selecting the articles. Furthermore, we devise an algorithm which iterates the above procedure until almost all records of experts' interest are selected. Experimental results by using the articles from Cell show that almost all records of experts' interest are selected while reducing the works of experts drastically. 1

We introduce a concept learning methodology for text understanding systems that is based on terminological knowledge representation and reasoning. Qualitybased metareasoning techniques allow for an incremental evaluation and selection of concept hypotheses. This methodology is particularly aimed at realworld text understanding environments where lexical /conceptual resources cannot be completely specified prior to text analysis and, as a consequence of partial understanding, competing concept hypotheses with different levels of credibility have to be managed. The Concept Acquisition Problem The work reported in this paper is part of a largescale project aiming at the development of a Germanlanguage text knowledge extraction system for two real-world domains --- information technology product reviews (approx. 100 documents with 10 5 words) and medical findings reports (approx. 120,000 documents with 10 7 words). The concept learning problem we face is two-fold. In the IT domain lex...

We introduce a methodology for knowledge acquisition and concept learning from texts that relies upon a quality-based model of terminological reasoning. Concept hypotheses which have been derived in the course of the text understanding process are assigned specific "quality labels" (indicating their significance, reliability, strength). Quality assessment of these hypotheses accounts for conceptual criteria referring to their given knowledge base context as well as linguistic indicators (grammatical constructions, discourse patterns), which led to their generation. We advocate a metareasoning approach which allows for the quality-based evaluation and a bootstrapping-style selection of alternative concept hypotheses as text understanding incrementally proceeds.

In this article we report our research on building FEA-TURES—an intelligent web search engine that is able to perform real-time adaptive feature (i.e., keyword) and document learning. Not only does FEATURES learn from the user’s document relevance feedback, but it also automatically extracts and suggests indexing keywords relevant to a search query and learns from the user’s keyword relevance feedback so that it is able to speed up its search process and to enhance its search performance. We design two efficient and mutual-benefiting learning algorithms that work concurrently, one for feature learning and the other for document learning. FEA-TURES employs these algorithms together with an internal index database and a real-time meta-searcher to perform adaptive real-time learning to find desired documents with as little relevance feedback from the user as possible. The architecture and performance of FEATURES are also discussed. 1.

Relevance feedback is a supervised learning method used to improve the effectiveness of information retrieval systems. It is an active learning technique, and may be the most widespread application of active learning to date. We briefly discuss information retrieval and the role of relevance feedback in it, and point out some areas where active learning could make further contributions. Information Retrieval The goal of information retrieval (IR) techniques is to provide content-based manipulation (retrieval, categorization, visualization, etc.) of large collections of objects. Often the objects are composed of natural language text, but IR approaches have also been applied to databases of stored speech, images, video, computer source code, and other forms of information. Conventional approaches to databases (e.g. relational databases) work well when the semantic of objects are well-understood and data processing can be reduced to working out the logical consequences of specified re...

In this paper we report our research on building FEATURES - an intelligent web search engine that is able to perform real-time adaptive feature (i.e., keyword) and document learning. Not only does FEATURES learn from the user's document relevance feedback, but also automatically extracts and suggests indexing keywords relevant to a search query and learns from the user's keyword relevance feedback so that it is able to speed up its search process and to enhance its search performance. We design two efficient and mutual-benefiting learning algorithms that work concurrently, one for feature learning and the other for document learning. FEATURES employs these algorithms together with an internal index database and a real-time meta-searcher so to perform adaptive real-time learning to find desired documents with as little relevance feedback from the user as possible. The architecture and performance of FEATURES are also discussed.