Affective computingSlide 2Computers which recognize emotions of the usersWhy would computers recognize emotion?Computers “faking” emotionsComputers “having” emotionsAn example of emotion: being afraidEEL 5708Affective computingLotzi BölöniEEL 5708Affective computing•Emotions and computers. •Initiator of the field: Rosalind Picard at MIT–http://affect.media.mit.edu/•Controversial at the first sight–Computers which have emotions?–Isn’t this too early (in the process of making computers intelligent)?–Commander Data: lots of intelligence – no emotions•In reality, the field deals with a lot of different fields, some of them not controversial at all.EEL 5708Computers which recognize emotions of the users•Essentially, it is a sensor problem.•We are very good at recognizing emotion when we are face to face–Maybe because it is important – but let’s leave this discussion for later•We have difficulty recognizing emotions through written text (except it is very well written). This is why we use smileys :-) :-( •Computers needs to recognize emotions from limited input:–Text typed–Typing speed, mouse movement (shaking hand etc)•But we can add extra sensors:–Blood pressure, heart rate, perspiration salinity, skin surface conductivity (sweaty hands etc).EEL 5708Why would computers recognize emotion?•Pilots, train conductors: should not get angry…•Security: recognize anxiety?•Lie detectors: used for long time•Interactive tutoring systems–detect boredom, interest, frustration–adapt•Detecting replicants–See “Blade Runner” / read “Do Androids Dream of Electric Sheep?” •No controversy here.EEL 5708Computers “faking” emotions•Basically a human-computer interface problem•You need to know what emotion to display, and when•Embodied agents, robots. •Applications:–Toys: virtual (Tamagochi), embodied (Aibo robotic dog)–Games: avatars in games–E-commerce systems: an affective virtual salesperson sales more (supposedly)–Interactive tutoring•Again, not controversialEEL 5708Computers “having” emotions•1. action selection (e.g., what to do next based on the current emotional state)•2. adaptation (e.g., short or long-term changes in behavior due to the emotional states)•3.social regulation (e.g., communicating or exchanging information with others via emotional expressions)•4. sensory integration (e.g., emotional filtering of data or blocking of integration)•5. alarm mechanisms (e.g., fast reflex-like reactions in critical situations that interrupt other processes)•6. motivation (e.g., creating motives as part of an emotional coping mechanism)•7. goal management (e.g., creation of new goals or reprioritization of existing ones)•8. learning (e.g., emotional evaluations as Q-values in reinforcement learning)•9. attentional focus (e.g., selection of data to be processed based on emotional evaluation)•10. memory control (e.g., emotional bias on memory access and retrieval as well as decay rate of memory items)•11. strategic processing (e.g., selection of different search strategies based on overall emotional state)•12. self model (e.g., emotions as representations of “what a situation is like for the agent")EEL 5708An example of emotion: being afraid1. afraid of spiders2. afraid of large vehicles3. afraid of a large vehicle careering towards you4. afraid of a thug asking you to hand over your wallet5. afraid your favorite party is going to lose the next election6. afraid you have some horrible disease7. afraid of growing old8. afraid that your recently published proof of Goldbach’s conjecture has some hidden
View Full Document