Search

The article discusses the use of recommender systems (RS) by various major companies. Topics include the role of recommender systems as part of daily online user experience and to support users in various domains, the tendency for recommender systems to predict the past rather than the future, and the impact of human-computer interaction on the success of an RS.

The article presents information related to the decision of the U.S. Supreme Court in the MGM v. Grokster case. In the Grokster case, MGM argued that technology developers should be held legally responsible for misdeeds of their users if they design a technology that facilitates wrongful acts, such as copyright infringement. The Supreme Court was not persuaded by the argument, and instead ruled that technologists who actively induce users to infringe are responsible for those infringements. Failure to install filtering software in a peer-to-peer (P2P) file-sharing technology was specifically mentioned as a possible factor by the Court, although insufficient for secondary liability. Litigation threats have caused some P2P developers to start redesigning their systems. Although MGM sprinkled its briefs with references to Grokster's failure to filter, there was no evidence before the Court about the feasibility or effectiveness of filtering technologies.

It is informed that surveys in the U.S. have found that 95 percent of schools, 43 percent of public libraries and 33 percent of teenagers' parents employ filtering software to block access to pornography and other inappropriate content. Many products are also now available to filter out spam email. In 2000, the U.S. Congress passed the Child Internet Protection Act mandating that schools and libraries install content-filtering software in order to be eligible for some forms of federal funding. A district court struck down the requirement for libraries on the grounds that it violates the First Amendment. Much of that court's finding of facts was devoted to analyses of error rates and some of the arguments made on appeal to the U.S. Supreme Court also hinged on analyses of error rates. Most empirical studies of error rates have suffered from methodological flaws in sample selection, classification procedures, or implementation of blocking tests. Results have also been interpreted inappropriately, in part because there are two independent measures of over-blocking that are sometimes confused, and likewise for underlocking. This article presents a framework to guide the design and interpretation of evaluation studies. While the framework applies with only minor modifications to the evaluation of spam filters, the examples and discussion here focus on pornography filters.

Mobile agents are changing the face of e-business and reshaping business models. In the process these agents are also posing new concerns regarding who really owns information. EBay Inc. made headlines three years ago by initiating a drastic new policy against third-party predatory search agents. The policy was directed against intelligent agents that would enter the auction site, search for items their issuers were looking for and then notify the issuer about pricing and deadlines. EBay's modified user agreement would prohibit third-party sites from collecting and sharing information found on eBay's site. The problem, as reported, was the search agents were frequently accessing eBay, sifting through auction offers, harvesting the information and placing it on alternate Web sites known as "aggregators." The list of aggregators included names such as biddersEdge.com, AuctionWatch. com, itrack.com and Ruby Lane. EBay claimed the aggregators' search agents were harmful in multiple ways. They would slow down eBay's transaction processing systems, thus reducing performance for all other eBay visitors. Moreover, outside search agents might not show the most up-to-date information and thus lower auction users' purchasing experience.

User-adaptive hypermedia systems are able to cater to users more effectively given the more information they possess about them. Adaptations that will be necessary are usually not known at the time when different pieces of information about users become available. Personalized hypermedia systems therefore have to lay this collected data in stock and maintain it for possible future usage. Data gathering is moreover mostly performed in an unobtrusive manner and often without users' awareness. This is done to avoid distracting users from their tasks and with consideration that users are very reluctant to perform actions not directed toward their immediate benefits, even when they would benefit in the long run. This current practice of data collection and processing in personalized hypermedia systems, specifically in personalized Web sites, seem to be in conflict with privacy concerns of Internet users. These became manifest in numerous recent consumer polls, where respondents reported being concerned about threats to their privacy when using the Internet, being concerned about divulging personal information online and being concerned about being tracked online. INSETS: OECD Basic Privacy Principles of National Application.;Some Countries With Restrictions on the Transborder Flow of....

The advent of Web browsers able to execute programs embedded in Web pages has enabled the use of embodied conversational characters on the Internet. In fact, an increasing number of Web sites employ animated characters for a large variety of tasks, virtual teachers guide users through online learning activities, Web chauffeurs provide recommendations on which links to follow and virtual insurance agents help users to fill in forms. Web-based sales agents do not just present the customer with various products, but try to make the visit to an online store a unique experience for an overview. Companies , such as Artificial Life, Extempore, Hate and Virtual Personalities, create virtual characters specifically for the Internet. The added value of an animated characters specifically for the Internet. The added value of an animated character is not just its ability to provide additional information about a Web site, as this could be done by other means just as well. More important is the creation of a new browsing experience for the user who no longer has to surf the Web on his or her own, but is instead situated in a social context with virtual characters.

This article discusses composing of crosscutting concerns using composition filters (CF). According to the author, certain design concerns, such as access control, synchronization, and object interactions cannot be expressed in object-oriented programming languages as a separate software module. These crosscutting concerns generally result in implementations scattered over multiple operations. The filters in CF model can express crosscutting concerns by modular and orthogonal enhancements to objects. Modularity means that filters have well-defined interfaces and are conceptually independent of the implementation (language) of the object. Orthogonality means that filter specifications do not refer to other filters. These two properties increase adaptability and reusability of concerns because filters can be applied to different languages and can be combined easily. In the CF model concerns are specified declaratively using a simple message manipulation language. This specification allows for various implementation strategies. Filter types encapsulate the accept and reject semantics of filters.

The article focuses on the privacy practices of web browser extensions. Do today's commercial free-of-charge software downloads show appropriate respect for the privacy of the computer user? One believes the answer to this question is no. The research group at the University of Denver Privacy Center addressed the question by investigating Microsoft's Internet Explorer (IE) browser extensions. These downloadable pieces of software improve IE by giving it the ability to automatically fill out Web forms, perform price comparisons while shopping online and liven up the interface with thematic images and sounds. A recent summary of browser extension products is available in. Browser extensions are usually free of charge-in exchange for clickstream and profile information about the user and access to the user's display for advertising. The spectrum of information practices among products is broad. While some require hill user address information and formal registration, others only request a zip code or age range at download time. Some products create a complete record of every Web site the user visits, while other products avoid performing any actions that could leave an audit trail hinting at the user's Internet activities and personal interests. INSET: Recommendations to Software Designers and Entrepreneurs..

The article deals with the issue of automatic personalization based on web usage mining. Most personalization systems for web fall into three major categories, manual decision rule systems, collaborative filtering systems, and content based filtering agents. The new generation of web personalization tools is attempting to incorporate techniques for pattern discovery from web usage data. Principal elements of web personalization of category include the modeling of web objects and subjects, their categorization, and determination of set of actions to be recommended. The discovery of patterns from usage data is primary in this, however, is not sufficient for performing personalization tasks. A variety of clustering techniques must be used in this analysis. The task of recommendation engine is to compute a recommendation set for the current user session, based on the user profile. It is here that web usage mining holds principal magnitude. This model of web personalization has applications even outside of electronic-commerce. INSETS: Data Preparation for Web Usage Mining.;Experiments with the WebPersonalize System.;Mining Association Rules for Personalization..

The article examines the transis approach, which is a unique large-scale multicast service designed for partitionable operation. Transis Project was designed with the following goals: Firstly to tackle network partitions and provide tools for recovery from them. Transis was designed to support partitionable operation, in which multiple network components that are disconnected from each other operate autonomously. Secondly, to meet the needs of a large network through a hierarchical communication structure, with gateways selectively filtering messages among domains. And lastly, to exploit the available network multicast within each local-area network and provide fast cluster communication. The Transis approach distinguishes itself in allowing partitionable operation and in supporting consistent merging upon recovery. It provides the required flexibility for building diverse fault-tolerant applications, some of which benefit from continued partitioned operation. Transis provides transport level multicast communication services.

Presents several letters to the editor about computer science, published in the April 1, 1993 issue of the journal "Communications of the ACM." Need for mathematical and technical foundations for computer science students at the baccalaureate level; Role of information retrieval and filtering in case-based reasoning; Suggestions for computer science and engineering students for achieving their goals.

The GroupLens project designed, implemented and evaluated a collaborative filtering system for Usenet news--a high-volume, high-turnover discussion list service on the Internet. Usenet newsgroups--the individual discussion lists--may carry hundreds of messages each day. While in theory, the newsgroup organization allows readers to select the content that most interests them, in practice, most newsgroups carry a wide enough spread of messages to make most individuals consider Usenet news to be a high noise information resource. Both taste and prior knowledge are major factors in evaluating news articles. The combination of high volume and personal taste made Usenet news a promising candidate for collaborative filtering. The authors determined the potential predictive utility for Usenet news was very high. This article discusses challenges involved in creating a collaborative filtering system for Usenet news.

Online readers are in need of tools to help them cope with the mass of content available to the World-Wide Web. In traditional media, readers are provided assistance in making selections. This includes both implicit assistance in the form of editorial oversight and explicit assistance in the form of recommendation services such as movie reviews and restaurant guides. The electronic medium offers new opportunities to create recommendation services, ones that adapt over time to track their evolving interest. Fab is such a recommendation system. This article explains how a hybrid system can incorporate the advantages of both methods while inheriting the disadvantages. The authors describes the two approaches for content-based and collaborative recommendation, explain how a hybrid system can be created, and then describe Fab, an implementation of such a system. Fab is a distributed implementation of a hybrid system, and is part of the Stanford University digital library project.

People Helping One Another Know Stuff (PHOAKS) is an experimental system that addresses this problem through a collaborative filtering approach. PHOAKS works by automatically recognizing, tallying, and redistributing recommendations of Web resources mined from Usenet news messages. PHOAKS consists of a general architecture for filtering information from electronic messages and a set of techniques for generating and managing dynamic Web-based interfaces. The PHOAKS system contains six months or more of recommendations and associated data for about 1,500 newsgroups. Thousands of new opinions about the Web resources are added weekly. PHOAKS architecture has been used to recognize and process Web resource recommendations and FAQ messages. The architecture consists of three main processes search, categorization and disposition.

The article focuses on open source processing frameworks and real-time computing systems. It comments on the Apache Hadoop software system which uses computer clusters to process large datasets and talks about how online music service Pandora uses Hadoop to analyze data on skipped songs and listener ratings through machine learning, collective intelligence tasks and collaborative filtering. It mentions search engine company Google's distributed data system, Google File System (GFS), and MapReduce which allows jobs to be broken into smaller pieces and sent to different computers. It states that Hadoop uses two software modules, one a file system similar to GFS which disperses large datasets on multiple computers, and Hadoop MapReduce which splits up and analyzes mined data.

The article discusses efforts of computer scientists to prevent computer users and consumers from receiving unsolicited electronic mail messages (spam). Spam is defined as disruptive, typically commercial messages that are sent as e-mail or posted on a computer network. Anti-spam researchers utilize sophisticated filtering and identification technologies to reduce the spam that reaches computer end-users. However, the costs of these efforts are borne directly by Internet service providers (ISPs) and indirectly by end-users when legitimate e-mails are classified as spam.

The article discusses the implications of filtering on democracy. As a result of the Internet and other technological developments, many people are increasingly engaged in the process of personalization, limiting their exposure to topics and points of view of their own choosing. They filter in and out, with unprecedented powers of precision. These developments make life much more convenient and in some ways much better; we all seek to reduce our exposure to uninvited noise and many of us like to read opinions we find congenial. But from the standpoint of democracy, filtering is a mixed blessing. Above all, in a heterogeneous society, such a system would require something other than free, or publicly unrestricted, individual choices. Emerging technologies, including the Internet, are hardly an enemy here. We can certainly use them to learn more, rather than to live in an echo chamber. But to the extent they weaken the power of general-interest intermediaries and increase our ability to wall ourselves off from topics and opinions we would prefer to avoid, they create serious dangers to democracy.

Simple Object Access Protocol (SOAP) is the primary transport mechanism for the Web services architecture. It is a lightweight protocol that can replace more complicated, distributed object technologies for many applications. However, use of Web servers, by SOAP to tunnel through firewalls, may limit its usefulness because it potentially opens corporations to external access. As it becomes more ubiquitous, SOAP may increasingly find itself in the sights of security personnel. SOAP is Based On Extensible Markup Language (XML) and Hypertext Transfer Protocol (HTTP). SOAP can tunnel through firewalls without any concessions from network security personnel. SOAP transmits messages by placing a routing engine (part of the SOAP processor) behind the Web server. As SOAP is increasingly used to access potentially powerful internal services, firewalls may filter its use. This effect has already been seen with email, as many organizations now strip potentially virus-laden attachments from all incoming and outgoing email.

Permission marketing requires consumers' consent before a Web site can track them with cookies, or send them marketing email, or sell their data to another company. Yet a study conducted by Cyber Dialogue Inc. found that 69% of U.S. Internet users did not know they had given their consent to be included on email distribution lists. Using the right combination of question framing and default answer, an online organization can almost guarantee it will get the consent of nearly every visitor to its site. Although lists of people who have supposedly opted-in for permission marketing schemes are valuable sources of revenue for Web sites, high response rates alone do not mean these lists contain valuable customers. They systematically explored the influence of question framing and response defaults on consumers' apparent privacy preferences in two online experiments detailed. Participants in these experiments were members of the Wharton Virtual Test Market, an online panel of over 30,000 Internet users representative of the U.S. Internet population. Results of experiments highlight the need for all online consumers to pay close attention to what they agree to when they send responses to a Web site. If consumers had fixed policies about the privacy of their data, then asking them to opt-out or opt-in to a Web site's privacy policy would make no difference to their answer.

The author reviews how the Chinese government has released two cartoons characters to assist in its effort to control what it's citizens can view on the Internet. According to the author, news of the cartoon characters named "Jingjing" and "Chacha" was first reported in "The Financial Times." It is suggested that the characters will be used to scare off anyone who attempts to subvert Internet laws. The author states that content considered subject by the Chinese government reportedly fell more than 60% after the release of the cartoons.

Security is often described as a weak-link phenomenon. A speaker described how a compiler could be modified to plant a Trojan horse into the system's login authentication program so that it would accept a known password. Twenty years later, efforts such as the Trusted Computing Group, Intel's LaGrande, and Microsoft's Next Generation Secure Computing Base are aiming to create secure systems from the ground up. "Trusted Computing" (TC) platforms include specialized hardware or a processor that can monitor a system's boot process to ensure the computer is based on appropriately certified hardware and software. Thus, a TC platform can be used to rigorously enforce third-party-mandated security policies such as those needed for digital rights management and mandatory access control. Recently a group managed to run Linux on a Microsoft Xbox without any hardware modifications. The Xbox, in common with many other game consoles, mobile phones, and even printer cartridges, can be considered an instance of a special-purpose TCplatform.

The next generation in web-based systems is coming closer as people move from the one-size-fits-all site to the customizable system that adapts to a user's individual needs and tastes based on the system's knowledge of the user. For many years an adaptable Web site was little more than a seemingly unachievable dream for most users. Today, with the intricate interweaving of such technologies as artificial intelligence, hypertext, machine learning, information retrieval and cognitive science, adaptive sites are emerging as the new face of Web-based systems. This month's special section of the periodical "Communications of the ACM" traces the road to the Adaptive Web. The author of this article is grateful that many leading visionaries in the field participated in this project, led by guest editors Peter Brusilovsky and Mark Maybury. Together they tell of systems that adapt their behavior to the goals, interests and tasks of individual users or groups of users. Their articles focus on adaptive interfaces, animated agents, personalization techniques and privacy concerns. They also provide a preview of promising areas of application and describe where future research is headed.

The Internet and World Wide Web may be the ultimate double-edged swords. They bring diverse opportunities and risks. Proposed legal sanctions, social pressures and technological means to prevent or limit access to what is considered detrimental all appear to be inadequate as well as full of risky side effects. Web self-rating is a popular notion, and is being promoted by the recent Internet Content Summit as an alternative to government regulation. The reliability of third party filtering is notoriously low. A popular commercial filter blocked the Privacy Forum when one of its raters equated discussions of cryptography social issues with prohibited "criminal skills." Sites may not know they're blocked and procedures for appealing blocking are typically unavailable or inadequate. There's also a concern that self-rating schemes will turn the Internet into a homogenized environment dominated exclusively by large commercial media operations.