Heriot-Watt University is leading research into creating artificial personalities that will react with people as humanly as possible, mirroring a scenario that is behind the latest Joaquin Phoenix movie 'Her'.
In the movie, a lonely writer develops a relationship with his newly purchased operating system that's designed to meet his every need. And while a fully responsive conversational computer system may not be reality yet, Heriot-Watt University and its partners are making the concept more credible year on year.
The applications we are developing concentrate on keeping humans at the centre of interactions. We are building systems that reflect the flow of human conversation and that can also learn from and become more attuned to their human interlocutors
Intelligent interactive systems
Professor Oliver Lemon, of the Interaction Lab in the University's School of Mathematical and Computer Sciences, said current research revolves around making systems and computers properly conversationally interactive with their human companions.
"Whether it is equipping a robot to assess and respond to multiple requests for attention, or having a virtual companion walk you through the streets of a strange city, answering your questions and helping you navigate, the potential for smarter computer systems is expanding.
"The concept of the knowing machine is set within popular culture, but rather than visions of HAL in 2001 A Space Odyssey, the applications we are developing concentrate on keeping humans at the centre of interactions. We are building systems that reflect the flow of human conversation and that can also learn from and become more attuned to their human interlocutors."
Heriot-Watt researchers are currently working on a several EC-funded research projects which all include elements of interactivity and conversational operation. Among them are:
- PARLANCE, which designs and builds mobile applications that approach human performance in spoken conversational interaction. It creates systems that can generate conversation in real-time, adapting to new concepts without manual intervention, and personalising spoken interaction.
- SpaceBook, a speech-driven, hands-free, eyes-free device for pedestrian navigation and exploration. Spacebook uses speech to interactively guide its users along pedestrian route plans and describe visible landmarks, as well as to answer questions about famous people and places.
- JAMES, a socially intelligent robot bartender who uses physical, visual, and spoken cues to recognise customers, figures out who needs attention, and then serves them in order.
Professor Lemon said, "Trying to build human-like capabilities for conversational interaction is an incredibly challenging problem, because it requires combining information from speech, vision, gestures, and facial expressions, as well as spatial location and movement. New mobile technologies are giving us the potential to access to huge volumes of data that we are using to create machines that can learn from their interactions with humans. Siri is just the tip of the iceberg."