<<Reader of PPT notes: this is intended to be a lighthearted talk about the possibilities of cyber-awareness through developmental psychology theory.>>
Many new terms more recently have come about such as artificial intelligence, android, cyborg, artificial general intelligence, etc. While tonight I may be referring to androids, I’m going to use the word ‘robot’ because of the original intention of the word, and also I just really like the word - I’m nerdy like that.
So a thing that has gotten my interest lately is how we are ‘teaching ‘ our robots to learn cognitively. And the theoretical possibility that by changing the way we teach robots from a developmental perspective, we can actually affect the possible outcome. So within that arises the question: can we use human development process to teach robots to be socially intelligent and do that via language?
Can we use human dev process to teach robots to be socially intelligent and do that via language? Can we use video augmented imitation to teach language and within that have a programmed context?
The word ‘robot’ came from the Czech word ‘robota’, which means ‘forced work’ or ‘slave labor’. It was first used in 1921 by a writer named Karel Capek in a play he wrote called R.U.R. (Rossum's Universal Robots). During his writing he was having a hard time naming his new fictional entities so his brother Josef the famous painter and writer suggested ‘roboti’.
Karel used the term ‘robot’ to refer to the machines that were simulating humans. The gist of the play was the robots were used as slave laborers, until they suddenly acquire souls. They then rebel against their creators and attempt to annihilate the human race.
Developmental Psychology is the psychology of humans as they age, in this case and in most cases in regards to young children. Dev Psych studies things like biological innateness or environmental influence, psychological characteristics that change over time and the factors that influence said change.
So two of the biggest schools of thought in Dev Psych are from Piaget and Vygotsky. Piaget was a cognitive constructionist, that is he felt that action originating from the self was what constructed our cognitive abilities. We call this psychogenesis. So, a baby’s curiousity plus time and experimentation equaled cognition.
Vygotsky however believed that sociogenesis, or cultural mediation, was how cognition was acquired. As a social constructivist he felt that mental functions where brought out by a child’s social interaction plus culture plus meaning derived from tool and signs equaled cognition.
So, for Piaget we would have his stages be a newborn touching a sensing things, learning in that way. Next would be a toddler learning to pick up and hold a cup to drink. Next would be finally understanding another individuals point of view. Next would be abstract logic understanding.
Vygotsky’s theories don’t preclude Piaget’s, but theres just the disagreement on where cognition comes from. So, like a baby listening to the speech patterns of the parents. And then environment ‘self-talk’ which becomes a tool for self directing behaviour. Which would later finally ‘inner speech’ that is thinking inside your head, which is something very young children cannot do.
So here is the crux of it – can we teach a robot to learn via a Vygostky framework? It would require teaching via imitation, creating an interaction dynamic by synchronizing movements via sensors. Here’s the interesting part - from there one would teach something called a ‘proto-language’ which for each movement there is a specific meaning like ‘I move my left arm’ or ‘I reach you’.
This would transition a language within the interaction. One study did this and actually resolved in an autonomous response from the robot. The robot through the interaction patterns became capable of the response through a numerical non terminating algorthithym. Via social interaction they were able to teach a robot to autonomously use this proto-language.
Language is a cultural invention. When we teach our robots the ability to communicate, which shows the ability of cognition, there is then the hypothetical recursive possibility for robots to then reach back and create a culture.
And here is the macro from the micro – culture begets language, language begets history (because you can speak of ‘I did pick up the cup’ and ‘I am picking up the cup’ . And with culture and language and history, you get…society.
A robot society! <wait for laughter> To me that’s a fascinating idea, robots being able to create their own society. There’s a nice picture of robots as part of the universal structure. So - am I basically saying that by socially interacting with robots, so that they are able to learn language, that they will eventually recursively bring about a robot society? My curious nature wants to believe.
Here are some really great articles that I pulled concepts from, I really highly suggest y’all to read them if you like, especially the work of Dr. Dautenhahn, and go ahead and email me if you would like for further conversation.
And who knows, next time we might be sitting next to these guys at <insert conference here>…
Psychogenesis v Sociogenesis
Psychogenesis v Sociogenesis: SocialCognitive Design of Robots By Sharon Greenfield
Can Humans Create a Robot Society?• Mimic human development process• Teach robots to learn to be socially intelligent• Robot culture recursively created via language
Cognitive Design of Robots• Most robots are designed without regards to social behaviour• Interaction dynamics matters; context via language• Use imitation to augment social interaction while engaged in robot language learning
Lil’ Bit O’ History Written in 1920 Karel Capek’s R.U.R.: (Rossums Universal Robots) Czech English ‘robota’ = ‘slave labor’
Lil’ Bit O’ History Written in 1920 Karel Capek’s R.U.R.: (Rossums Universal Robots) Czech English ‘robota’ = ‘slave labor’ 1. Acquire soul 2. Rebel against ‘owner’ 3. Attempt annihilation of said ‘owner’
Developmental Psychology What is it?• Lifespan physical, cognitive, and social change Why does it matter?• Nature/Nurture• Process• Change
Psychogenesis v SociogenesisLev S. Vygotsky = Social Constructivism + + =
Piaget Framework Emphasis:1. Sensorimotor period (Ages 0-2yrs): Learn via senses, touch, movement. Very egocentric.2. Preoperational period (Ages 2-7) : Better motor skills, object orientation. Very intuitive.3. Concrete operations period (Ages 7-12) : Large logical leaps ability. No longer egocentric.4. Formal operations period (Ages 12-15): Ability of logic, abstract reasoning, hypothetical thought.
Vygotsky Framework Emphasis:• Cognitive development changes per cultural and historical factors• Acquire cultural tools via social interactions• Symbolic knowledge acquisition• Language is key to cognitive development
Gah Gah Gah Goo… Using the Vygotsky framework one might:• Socially interact with robot via imitation• Synchronization, create interaction dynamic• Teach a proto-language; per each movement, a meaning/specific ‘word’.
Gah Gah Gah Goo… Using the Vygotsky framework reply could be:• Study language transition• Track the autonomous response• Track the robot-human interaction patterns• No dirty diapers
References and Thank You Thanks to Aubrey and Ren for consultation, to all the hardworking Ignite organizers and volunteers, and to the Portland community for being interested! -Sharon human.factor.one@ gmail.com