Robotics Reference
In-Depth Information
the Cathexis model of emotion created by Juan Velazquez. Expectations
change over time, for example a person may expect to pass an exam but
after taking a few courses and failing the exam each time his expectation
of passing the exam next time will be much lower. Therefore it is impor-
tant to allow the robot's expectations to change with its experiences.
Empathy—How Robots Recognize and Measure
Emotions in Humans
The most satisfactory personal interactions are those in which everyone
involved in the interaction can not only express their own emotions but
also recognize the emotions of others. An important aspect of relation-
ships between humans is empathy—the ability to imagine ourselves in
the position of another person, to experience what another person is
feeling. Empathy gives us access to a person's mental states, to his or
her desires, emotions and beliefs, and thereby plays an indispensable role
in our social interaction and communication.
Empathy is one of the essential ingredients for long-term success at
the poker table. In poker it is important to recognize “tells”, those sub-
conscious gestures that give away information to the opponents. Stuart
Marquis and Clark Elliott have developed a set of poker playing programs
at DePaul University that use emotion in two ways. These programs dis-
play emotion
6
and they recognize emotion in their human opponents.
The control program uses voice recognition software,
7
appliedtothe
inflection in an opponent's voice, to distinguish between seven broad
emotion categories: neutral, joy, fear, anger, hate, sad and love. This
classification enabled the programs to distinguish the limited essential
vocabulary of poker (“bet”, “check”, “call” and “raise”) according to the
inflection of the opponent's voice. The classification here is into seven
categories rather than the 22 basic emotions in the Ortony/Clore/Collins
model, because the current state of the art in voice recognition does
not provide sufficient discrimination to cope with so many gradations
of voice. As hardware speeds and memory sizes burgeon in the fu-
ture, more accurate voice recognition will be possible, enabling robots
6
See “How Robots Express Emotion” later in this chapter.
7
It is important to distinguish between Speech Recognition, which is the science of recognizing
what words are spoken, and Voice Recognition, in which the task of the technology is to recognize
the voice of a particular person or to determine that person's mood or stress level from characteristics
of their voice.
