Tag Archives: signal processing

Yikes! New Computer System Can Read Your Emotions

28 Nov

Researchers from the Universidad Carlos III de Madrid has developed an emotion-reading AI.

It’s bad enough your significant other insists that there’s something wrong when there really isn’t, but now it looks as if we’ll be dealing with an emotion-reading AI in the not-too-distant. But don’t get too frightened just yet: we may still be a ways off from talking to HAL 9000 and listening to him/it cheering us up first thing in the morning with a little tune if we seem a little down in the dumps. 

The study was published in the Journal on Advances in Signal Processing. As explained in their study, the computer system created by these researchers independently adapts its dialogue to the users’ situation so that its responses are on the same page as the users’ emotional state. The system uses up to 60 different types of acoustic parameters including tone of voice, speed of speech, duration of pauses and even the energy of the voice signal. Significantly, it looks for negative emotions like anger, boredom and doubt.

When the emotion is detected, the system then finds out the user’s overall intention in a given dialogue. A news release by the Universidad Carlos III de Madrid states an example.

For example, if the system did not correctly recognize what the interlocutor wanted to say several times, or if it asked the user to repeat information that s/he had already given, these factors could anger or bore the user when s/he was interacting with the system. Moreover, the authors of the study point out that it is important that the machine be able to predict how the rest of the dialogue is going to continue.

The researchers solved the problem by developing a statistical method that uses earlier dialogues to learn what actions the user is most likely to take at any given moment. Once the system has detected the user’s intention — along with his/her emotional state — it automatically adapts the dialogue to the situation the user is experiencing. The report reads,

For example, if s/he has doubts, more detailed help can be offered, whereas if s/he is bored, such an offer could be counterproductive. The authors defined the guidelines for obtaining this adaptation by carrying out an empirical evaluation with actual users; in this way they were able to demonstrate that an adaptable system works better in objective terms (for example, it produces shorter and more successful dialogues) and it was perceived as being more useful by the users.

There is no doubt, if this AI goes commercial, we might as well kiss human tech support — whether it’s local or based overseas.

Source: tomshardware