Abstract: A facility for generating item recommendations is described. In one embodiment, the facility displays a plurality of item indications each indicating an item. The facility then receives user input selecting one or more of the displayed item indications, and selects as seed items the items indicated by the selected item indications. The facility then uses the selected seed items to generate a list of recommended items, and displays the generated list of recommended items.

Abstract: Just prior to a scheduled session with a live instructor, a student is presented with an automated lesson which prepares him for the live instruction. The automatic lesson presents and practices skills that will be needed for the live lesson. Thus, when the instructor appears, the student is entirely prepared for the live lesson and optimum use may be made if the instructor's time. After the live lesson is complete, the student may be returned to an automated debriefing, which helps imprint major aspects of the live lesson of the students of mine. Thus, the learning process is improved through the combination of live and automatic instruction.

Abstract: A facility for generating item recommendations is described. In one embodiment, the facility displays a plurality of item indications each indicating an item. The facility then receives user input selecting one or more of the displayed item indications, and selects as seed items the items indicated by the selected item indications. The facility then uses the selected seed items to generate a list of recommended items, and displays the generated list of recommended items.

Abstract: A facility for generating item recommendations is described. In one embodiment, the facility displays a plurality of item indications each indicating an item. The facility then receives user input selecting one or more of the displayed item indications, and selects as seed items the items indicated by the selected item indications. The facility then uses the selected seed items to generate a list of recommended items, and displays the generated list of recommended items.

Abstract: The present invention is directed to automating the sending of campaign messages in computing environments. Campaign scores are determined for messaging campaigns and eligible recipients are assigned to campaigns with a higher score. More specifically, eligible recipients are assigned to a particular messaging campaign if the campaign is the “best” campaign for each eligible recipient based on the campaign score. Additionally, a portion of the eligible recipients are reserved as a control group for each campaign. After campaign messages have been sent out to assigned recipients, the behavior of the recipients who received campaign messages and the behavior of the control group members are differentiated to update the campaign's score. Continually updating campaign scores makes it possible to continually update which campaign is best for a particular recipient.

Abstract: The present invention is directed to automating the sending of campaign messages in computing environments. Campaign scores are determined for messaging campaigns and eligible recipients are assigned to campaigns with a higher score. More specifically, eligible recipients are assigned to a particular messaging campaign if the campaign is the “best” campaign for each eligible recipient based on the campaign score. Additionally, a portion of the eligible recipients are reserved as a control group for each campaign. After campaign messages have been sent out to assigned recipients, the behavior of the recipients who received campaign messages and the behavior of the control group members are differentiated to update the campaign's score. Continually updating campaign scores makes it possible to continually update which campaign is best for a particular recipient.