#raw eeg

This is an Emoshape Affective Toy robotic doll you can construct in 10 minutes that understands your emotions. 

This is a totally wired up person having all their thermal image vital signs registered in Dolby Labs. This does not take 10 minutes.

2018 is going to be the year that biometrics, robotics and consumer technology are going meet and greet all over the place.

Emoshape

Emoshape, hails itself as ‘emotions next frontier’ uses a chip with tons of processing power to work with language to generate EPU (emotional processing units). It uses a patent pending technology that creates emotional states and synthetic emotion in intelligent machines. It claims to distinguish between twelve primary emotions:  anger, fear, sadness, disgust, indifference, regret, surprise, anticipation, trust, confidence, desire and joy. The emotion recognition classifiers says it achieves up to 94 percent accuracy on conversation. It does this with a microcontroller using real-time appraisal computation with reinforcement learning.  An Emotional Profile Graph (EPG) computation functionality allows the AI or robot to experience 64 trillion possible distinct emotional states. Emotional stimuli is stored within the memory bank through emotional patterns or fingerprints.

64 trillion emotional states? Like I don’t have enough to worry about? What does having 64 trillion emotional states actually feel like? It exhausts me just thinking about it. 

12 emotions graphed in realtime - great representations for the synesthesiac crowd.

How the colors are assigned - although I beg to synesthesiacally  differ

It uses “sound waves of our consciousness” (not sure about that one). To drill down further, the Emoshape has two recognition layers that work at the sentence level and uses an form or evolution of the Ekman emotion classification, first developed in 1993. From what I have seen a lot of people are now redefining that classification, but it does give a basic structure to operate within. It then incorporates a  semantic level that refers back to the Emoshape API cloud service  based on Patrick Levy-Rosenthal psychobiotic evolutionary theory. I wondered what that was, so I found an article from Trends In Neuroscience which said, “Psychobiotics were previously defined as live bacteria (probiotics) which, when ingested, confer mental health benefits through interactions with commensal gut bacteria.” Interesting. The Neuroscience article goes on to say “Much psychobiotic research is based on rodent models.”

Original Chatty Cathy Mattel dolls and their wonderful wardrobes

Pull the string - and she says 11 different things!

The buffers of how the emotional stacks are weighted and assembled in Emoshape. 

However, this part looks quite complex and well thought out. And the flow chart below absolutely amps it up.

But still, it all seems semantically based, like a huge taxonomy of words though there is an API for physical interaction, and a psychophysics API, which at this stage seems to mean a relationship between mental phenomenon and physical stimuli - a pretty wide net for sure .

Dolby Labs

Dolby Labs has been around since 1965, but they now place a high premium on people watching movies. This is so they can show their clients that certain colors, sounds, or frequency ranges will elicit a certain response. Neurophysiologist Poppy Crum run the labs. Affective computing is not new news, but the entertainment industry is now all over this one. Netflix and Hulu at this point use eyetrackers.

Poppy Crum at Dolby Labs

But what do these biosensors really talk about? Well, to begin with a subject is outfitted with a 10/20 brain EEG Cap.

Brainwave EEG headset being fitted on TV host Lauren Goode

 Muscle detector and lie detector also being booted up. Thermal images are also captured (see beginning of post)

Raw EEG, GSR and heart rate are all displayed livetime.

Some of the things they discovered that if you are shown a huge screen with flames, your body begins to flush - and that apparently helps shape the stories.

As immersive environments or ‘mixed reality’ grows in influence, it is logical that tracking the emotional responses of users needs to move along as well.

Disney

Disney is now so all over affective emotions and measurements they have teamed with Cal Tech to start using Artificial Intelligence. They work with what they call  “factorized variational autoencoders,” or FVAEs, which after observing an audience member’s face for just 10 minutes,  can predict how a person will react to the rest of the movie. 

It can actually work on any type of time durational data - "We are all awash in data, so it is critical to find techniques that discover patterns automatically," said Markus Gross, vice president at Disney Research. "Our research shows that deep learning techniques, which use neural networks and have revolutionized the field of artificial intelligence, are effective at reducing data while capturing its hidden patterns."

Singing Christmas carols - can you pick out the unhappy one? Bonus points - Disney Research is based in.....Pittsburg - with free research trips to Disney Park