Abstract

Existing systems are able to learn information agents through demonstration that provide programmatic access to web-based information. However it is still difficult for end users to combine these information agents in procedures that are customized to their particular needs. We combine learning by demonstration with learning by instruction to build a system to learn such procedures with a small amount of human input. The instruction system relies on knowing the input-output types of the information agents in order to combine them. We make use a system that learns to predict the types from examples to simplify this part of the task. Our instruction system performs a search that has interesting similarities with proof search in explanationbased learning.
Intelligent software agents aim to assist users in the office environment by carrying out complex everyday tasks, for example planning travel, or managing the purchasing of equipment. These tasks are often on-going processes, where the assistant should initially combine information from a variety of heterogeneous sources and process the information as the user wishes, perhaps monitoring the progress of the task and continuing to give help where appropriate. For example, in planning a trip, the assistant may initially gather information about flights and hotels based on the user’s preferences and budget, make some recommendations and assist with booking the user’s choice, send reminders to the travelers, monitor for changes in prices, and check in for flights at the appropriate time. The software assistant must be able to learn new tasks and procedures, due to the wide range of potential tasks the …

Date

December 3, 2025

Authors

James Blythe, Dipsy Kapoor, C Knoblock, Kristina Lerman, Steven Minton

Journal

Proc. of the AAAI 2007 Workshop on Learning Planning Knowledge