Nobody likes having to deal with automated telephone services, that say wonderful things like, "You said 'Beelzebub,' is that correct?". Such services may get slightly less annoying, however, thanks to research being carried out at Spain's Universidad Carlos III de Madrid and Universidad de Granada. A team of scientists from those institutions have created a computer system that is able to recognize the emotional state of a person speaking to it, so that it can alter its behavior to make things less stressful.
The system analyzes a total of 60 acoustic parameters of users' voices, including tone, speed of speech, duration of pauses and energy of voice signal. The scientists designed the system to look for negative emotions in particular, that would indicate anger, boredom, or doubt.
Not only does the system draw its conclusions from what users' voices sound like, but also from how their conversation with it progresses. If the system is repeatedly not recognizing what a person is saying, for instance, or has to ask them to repeat information previously given, it's a likely bet that the person is getting angry and/or bored. Based on a statistical method gleaned from previous conversations, the system can also guess where a conversation with a user is heading, and what actions they're likely to take.
After having identified a person's mood and intentions, the system could then adapt the dialogue accordingly. If a user sounded doubtful of the system, for instance, it could offer them more help. If they sounded bored or angry, however, that offer might just irk them further.
The Madrid/Granada scientists tested a prototype system on human test subjects, and found that it resulted in shorter, more successful conversations. Down the road, perhaps it might someday be combined with a system being developed at Binghamton University, that identifies computer users' emotional states by looking at their faces.
