PRICAI 2002: Trends in Artificial Intelligence

Consumers and businesses have access to vast stores of information on the Internet, ranging from newspapers, shopping catalogs, restaurant guides, classified ads, jobs listings, dating services to discussion groups and e-mail. All this information is typically accessible only while users are in front of a computer at home or in an office. Wireless devices allow unprecedented access to information from any location at any time. The presentation of this information must be tailored to the constraints of mobile devices. Small screens, slower connections, high latency and limited input capabilities present new challenges. Agents that learn user’s preferences and select information for the user are a convenience when displaying information on a 19-inch desktop monitor accessed over a broadband connection; they are essential on a handheld wireless device. This paper summarizes commercially deployed systems using machine learning methods for personalizing mobile information delivery. 1 The Need for Adaptive Personalization Many wireless data services cope with small screens and limited capabilities of wireless devices by reducing the amount of information available to users. One news site makes four business articles available to mobile users and one wireless carrier limits the user to less than ten wireless games to download. This is equivalent to a newspaper having only the front page of each section available or an electronic game store having a single shelf of products. However, it is rarely the case that a small offering can satisfy the needs of a diverse group of consumers. Here, we describe the experience that AdaptiveInfo has had with personalizing delivery of wireless news and downloadable content such as Java applications for mobile phones. The experience with wireless readers of the Los Angeles Times wireless web site is that users read much more than what appears on the front page of the print edition – only 31.6% of the content read wirelessly is among the top five stories in each section on the regular web site, showing that mobile users want more content than a one-size-fits-all approach can deliver. Similarly, in an analysis of downloadable content for mobile personal digital assistants, we found that the top 10 items account for less than 40% of the sales for games and less than 30% of the sales for utilities. While a limited one-size-fits-all approach is easy to use, if a mobile