Robotics PDF

Reminders: Sign in to AttendanceRadar Quiz:  Robotics and Embodied Cognition  How does cognition change when situated in a physical environment? How should we design cognitive agents that exist in physical environments?  What problems require cognition or intelligence? Ø Acti...

Reminders: Sign in to AttendanceRadar Quiz:  Robotics and Embodied Cognition  How does cognition change when situated in a physical environment? How should we design cognitive agents that exist in physical environments?  What problems require cognition or intelligence? Ø Acting in the physical world feels easy for humans, but is surprisingly hard for machines Ø Moravec’s paradox: sensorimotor and perception skills require enormously more computational resources than abstract reasoning “The deliberate process we call reasoning is, I believe, the thinnest veneer of human thought … we are all prodigious olympians in perceptual and motor areas, so good that we make the difficult look easy.” –Hans Moravec Embodied Cognition Ø Traditional cognitive approach: think of the (natural or artificial) mind as separate from the environment, which provides inputs and outputs Ø Embodied cognition: think of the mind + environment as a single cognitive system Embodied Cognition Ø Refers to a whole spectrum of ideas, from: Ø “Simple” embodied cognition: our cognitive systems are made to operate in a physical world Ø “Medium” embodied cognition: states of the world and our bodies shape our thoughts in a fundamental way Ø “Radical” embodied cognition: The mind cannot be meaningfully studied in isolation from the world Modal representations Ø One branch of embodied cognition argues that most/all of our semantic representations are modal: they are tied to the sensory and motor information associated with that concept Ø To study this in humans, we can look for times when this kind of embodied information is present even when not necessary for a task Handle side Tucker & Ellis, 1998  1 Today 4 2 Where is yesterday? 3 Earlier Later or Later Earlier Fuhrman & Boroditsky, 2010  Reaction time (ms) Button order effect (ms) Fuhrman & Boroditsky, 2010 Why should we think of the environment as part of a cognitive system? Program a robot to: Ø Move forward whenever possible Ø If stuck, back up a bit and turn a random amount What task is this robot accomplishing? Program a robot to: Ø Move forward whenever possible Ø If stuck, back up a bit and turn a random amount What task is this robot accomplishing? Ø In this specific environment (with this specific robot body, blocks of this specific size), this robot has the function of helping to pile up blocks Ø If obstacle in left rear or right front, turn left Front Ø If obstacle in right rear or left front, turn right Right Left What does this robot do? Rear Ø If obstacle in left rear or right front, turn left Front Ø If obstacle in right rear or left front, turn right Right Left What does this robot do? Ø In a maze, this robot Rear will follow walls to solve a maze Ø If we want to study this system at Marr’s highest (functional/computational) level, it is only possible when considering the robot + environment Ø Consistent with “radical” embodied cognition: we can’t study these cognitive agents in isolation Off-loading cognition Ø These examples show that we can get away with very limited cognition on-board the robots Ø Humans can use their environment strategically in order to carry out cognitive processes Australian Aboriginal memory technique: associate information with physical locations, then walk back through these locations to remember Reser et al., 2021 Dynamical Coupling Ø Embodied cognition can be an ongoing interplay between an agent and the environment Ø As opposed to a “representation-hungry” system that maintains complex internal states and computes offline, systems can repeatedly make simple computations by observing the world Dynamical Coupling Ø Strategies for catching a baseball: Ø Measure and represent its position and velocity (and possibly its spin, the wind speed, etc.) and run a physics simulation algorithm to determine where it will land Ø Constantly adjust your running direction such that the ball looks like it is always moving up in the same direction McBeath et al., 1995 Rodney Brooks Ø “Actionist” or “situated” or “behavioral” robotics: argues against maintaining representational states inside robots Ø Rather than creating detailed internal models and making multi-step plans, robots should be composed of many interacting behavioral units that react to the environment Ø “the world is its own best model--always exactly up to date and complete in every detail” Brooks, 1991 Social cognition Ø We can also think of other people/robots as part of the environment in which we are situated Ø For some kinds of collective behaviors, we may need to think about the “basic unit” of cognition as consisting of groups of minds Program a robot to: Ø Increase a counter by 1, 30x a second Ø When counter reaches 100, flash a light and reset the counter to 0 Ø If you see a flash nearby, add 10% to your counter What is this robot computing? Ø This algorithm provides a way for a group to come to a consensus synchronization! ØThe “cognitive system” here is the whole group, not the individual Program a robot to: Ø Turn to align its direction with the average direction of nearby robots, and Ø Turn toward the average position of nearby robots What is this robot computing? https://www.harmendeweerd.nl/boids/ Autonomous driving requires modeling decision-making at the multi-car level Schwarting et al., 2019 Morality for robots Ø Since robots can take actions in the physical world, they have the potential to cause or prevent physical harm to people, animals, property, etc. Ø There are cases in which they may need to make a fast decision without time for confirmation by a human operator Asimov’s Laws of Robotics First Law: A robot may not injure a human being or, through inaction, allow a human being to come to harm Second Law: A robot must obey the orders given it by human beings except where such orders would conflict with the First Law. Third Law: A robot must protect its own existence as long as such protection does not conflict with the First or Second Law. Summary Cognitive systems that act in the physical world: Face massive computational challenges in carrying “simple” motor actions Can have abstract representations that are shaped by their physical interactions May be tightly coupled to their environments, with parts of their representations and algorithms extending into the external world Reminders Short paper #2 due Tuesday Nov 19th