From Answering Engines to Task-completion Engines

Web search engines over the past decade have evolved into being the primary gateways to accessing the ever-growing amount of data available online. Major web search engines (Google, Bing, and Yahoo!) have become extremely effective in responding to a range of requests directly and appropriately (e.g., by showing results on a map or displaying “info-boxes” for entities, such as people or organizations). Search, however, is rarely performed for its own sake, but is usually associated with a specific target or goal. In many cases, this goal is the completion of a larger task, which is often complex (involving a nontrivial sequence of steps) and knowledge-intensive (requiring access to and manipulation of large quantities of information). Planning a family vacation or setting up a task force are just two of a plethora of examples. Such tasks call for a potentially large number of search queries to be issued in order to collect all the information needed. And, it often takes additional data processing steps (filtering, sorting, aggregating) before an actionable decision can be reached. Contemporary search environments are tailored to support a small set of basic search tasks and provide limited help in this tedious process. Resolving complex tasks with current search technology often requires us to use multiple search sessions and multiple search strategies, and then manually synthesize and integrate information across sessions (i.e., opening multiple windows or tabs and cutting-and-pasting information between them). To solve these problems, one needs a paradigm shift from answering engines to task-completion engines.

The aim of this project is to develop, implement, and test a task-completion engine that supports humans in solving complex, knowledge-intensive tasks, by providing an integrated environment that caters for all task-related activities (which, to date, are performed using a combination of various tools, applications, and services). Our system will provide assistance for formulating information needs by engaging in a dialog with the user, will offer rich interaction with results, and will be able to learn from past user interactions. Put simply, it is about integrating features from Google, Excel, and SIRI into a single product.

See the SCST@ECIR’15 paper Task-completion Engines: A Vision with a Plan for a more detailed description of the project. This paper is also accompanied by a poster.

Excerpt from the envisaged user interface

Key components and their dependencies