A personal news agent that talks, learns and explains

Abstract

Most work on intelligent information agents has thus far focused on systems that are accessible through the World Wide Web. As demanding schedules prohibit people from continuous access to their computers, there is a clear demand for information systems that do not require workstation access or graphical user interfaces. We present a personal news agent that is designed to become part of an intelligent, IP-enabled radio, which uses synthesized speech to read news stories to a user. Based on voice feedback from the user, the system automatically adapts to the user's preferences and interests. In addition to time-coded feedback, we explore two components of the system that facilitate the automated induction of accurate interest profiles. First, we motivate the use of a multistrategy machine learning approach that allows for the induction of user models that consist of separate models for long-term and short-term interests. Second, we investigate the use of "concept feedback", a novel form of user feedback that is based on our agent's capability to construct explanations for the reasons that have led to a specific classification. Users can then critique these explanations which, from a machine learning perspective, allows for more direct changes to an induced concept than through the inclusion of additional training examples. We evaluate the proposed algorithms on user data collected with a prototype of our system, and assess the performance contributions of the system's individual components.

Ubiquitous User Assistance in a Tourist Information Server

“…To address this issue, we have extended the system to support logging without registering: if the user logs in this way, she receives non personalized recommendations, but she can access the other facilities offered by the system, such as category and geographical search, itinerary scheduling and presentation of tourist attractions. Moreover, we will investigate the application of unobtrusive user modeling techniques aimed at the identification of the user's preferences on the basis of the observation of her browsing behavior [6,9].…”

Buzzer – Online Real-Time Topical News Article and Source Recommender

“…Perhaps the earliest example of a news recommendation service, Krakatoa Chronicle [4], represented user profiles as a weighted vector of terms drawn from the articles that a given user liked, and matched this weighted vector against a new set of articles to produce a ranked list for presentation to the user. Similarly, Billsus and Pazzani's News-Dude [1] harnessed content based representations and multi-strategy learning techniques to generate short-term and long-term user profiles, as the basis for news recommendation. Although Billsus and Pazzani [7] argue that content-based approaches to finding trends and topics in news articles are difficult because of the sheer random bag-of-words unstructured nature of articles, and the complexity of natural-language processing.…”

“…To calculate an overall score for each article we simply compute the sum of the TF-IDF scores across all of the terms associated with that article as per Equation 1. In this way, articles which contain many tweet terms with high TF-IDF scores are preferred to articles that contain fewer tweet terms with lower TF-IDF scores.…”

“…There is a long history of using recommender systems techniques to help users to navigate through the myriad of news stories that are written and published everyday [1,2,6]. These systems can promote the most relevant stories to a user based on their learned or stated preferences or their previous news consumption histories, helping the user in question to keep up-to-date and to save valuable time sifting through less relevant stories.…”

Complex adaptive filtering user profile using graphical models

“…(Billsus & Pazzani, 1999) built a personal news agent that used time-coded feedback from the user to learn a user profile. However their way of using time as feedback is rather heuristic.…”