European researchers developing next generation search engine
A European research consortium recently got a €2.8 million boost from the EU’s research funding programme to develop a search engine that responds to questions asked orally. Once fully developed, the service will be multilingual with a focus on mobile applications so users can ask for and receive information on their mobile phones.
The main objective of QALL-ME (Question Answering Learning Technologies in a Multilingual and Multimodal Environment) is to develop technology that can extract the exact answer to a question rather than produce a list of thousands of documents that may or may not include the answer.
|Voice activated searches will be able to beam a map to the nearest subway station onto your mobile.|
Researchers are investigating the possibility of multilingual and multimodal open domain technology. If successful, users will be able to ask a question in one language and receive the answer in another, or ask for directions via a text message and receive a map containing their destination.
QALL-ME was conceived by the University of Wolverhampton Professor Ruslan Mitkov, Senior Lecturer Constantin Orasan, and their research group.
“In contrast to the technologies behind today’s
web search engines, the goal of QALL-ME is not to return hundreds
of documents, but the actual sequence of words which constitute
the answer,” Professor Mitkov said.
“We are particularly targeting users of the latest mobile
phones who want accurate information they can act on instantly.
It’s all about making services more user-friendly and
obtaining information more conveniently.”
The principle technology behind the project, known as Open Domain Question Answering (QA), takes a question in natural language and produces an answer from various information sources (e.g. documents, databases, etc.)
The research objectives of the project aim to move away from current factoid based questions that are usually associated with voice portals. Researchers hope to integrate temporal and spatial contexts both for question interpretation and for answer extraction into the technology.
For example, the answer to “When is the next bus from
the train station to the airport?” depends on where you
are and when you ask it. Mr Mitkov and colleagues hope to develop
the technology that takes such factors into consideration so
that you receive the correct answer regardless of whether you
are in Brussels or Bangkok.
The project is planning on using an open source architecture so any number of companies will be able to use it in the future. The project will last for three years and will receive more than half of its funding from the Commission.
IST Project Fact Sheet
The Research Institute in Information and Language Processing (RIILP)