Intro-Cogmod

Cognitive Modeling(Brain -> Mind) : 

Cognitive Modeling(Brain -> Mind) Psyc 4510, Cogs 6964, CSCI 4964

“How do brain processes give rise to mind?” J. McClelland“How does the brain think?” O’Reilly & Munakata : 

“How can the mind exist in the physical universe?” John Anderson “How do brain processes give rise to mind?” J. McClelland“How does the brain think?” O’Reilly & Munakata

Reductionism : 

Reductionism Physical Reductionism The brain Cognitive Neuroscience Neurological Components Cognitive Architectures Brain Areas

ACT-R 6.0 Mapping to the Brain* : 

ACT-R 6.0 Mapping to the Brain* Environment Productions (Basal Ganglia) Retrieval Buffer (VLPFC) Matching (Striatum) Selection (Pallidum) Execution (Thalamus) Goal Buffer (DLPFC) Visual Buffer (Parietal) Manual Buffer (Motor) Manual Module (Motor/Cerebellum) Visual Module (Occipital/etc) Intentional Module (not identified) Declarative Module (Temporal/Hippocampus)

Reconstructionism : 

Reconstructionism Brain is made of neurons Not enough, need explanation How billions of interacting neurons produce (embodied) Cognition Need computational approach Simulations to explain complex phenomenon

Emergent Phenomena : 

Emergent Phenomena Arise from interactions with obviously being present in the behavior of the individual units

Levels of Analysis : 

Levels of Analysis Brain Billions of neurons, 5000 x billion synapses Dynamic Marr’s Levels of Abstraction Computational Algorithmic Implementation

Levels of Analysis : 

Levels of Analysis Focus only on 1st two levels? Assume a particular implementation Focus only on third level Which details are important?

Leabra 2 level approach : 

Leabra 2 level approach Cognitive Phenomena Neurobiological Mechanisms Principles

Newell’s Time Scale of Human Activity (amended) : 

Newell’s Time Scale of Human Activity (amended)

What is a Cognitive Architecture? : 

What is a Cognitive Architecture? Infrastructure for an intelligent system Cognitive functions that are constant over time and across different task domains Analogous to a building, car, or computer

Integrated Cognitive Architecture : 

Integrated Cognitive Architecture Cognition does not function in isolation Interaction with perception, motor, auditory, etc. systems Embodied cognition Represents a shift from “mind as an abstract information processing system” Perceptual and motor are merely input and output systems Must consider the role of the environment Other body processes Effects of caffeine, stress and other moderators

Motivations for a Cognitive Architecture * : 

Motivations for a Cognitive Architecture * 1. Philosophy: Provide a unified understanding of the mind. 2. Psychology: Account for experimental data. 3. Education: Provide cognitive models for intelligent tutoring systems and other learning environments. 4. Human Computer Interaction: Evaluate artifacts and help in their design. 5. Computer Generated Forces: Provide cognitive agents to inhabit training environments and games. 6. Neuroscience: Provide a framework for interpreting data from brain imaging. 7. All of the above

Requirements for Cognitive Architectures* : 

Requirements for Cognitive Architectures* 1. Integration, not just of different aspects of higher level cognition but of cognition, perception, and action. 2. Systems that run in real time. 3. Robust behavior in the face of error, the unexpected, and the unknown. 4. Parameter-free predictions of behavior. 5. Learning.

History of the ACT-framework* : 

History of the ACT-framework* Predecessor HAM (Anderson & Bower 1973) Theory versions ACT-E (Anderson, 1976) ACT* (Anderson, 1978) ACT-R (Anderson, 1993) ACT-R 4.0 (Anderson & Lebiere, 1998) ACT-R 5.0 (Anderson & Lebiere, 2001) Implementations GRAPES (Sauers & Farrell, 1982) PUPS (Anderson & Thompson, 1989) ACT-R 2.0 (Lebiere & Kushmerick, 1993) ACT-R 3.0 (Lebiere, 1995) ACT-R 4.0 (Lebiere, 1998) ACT-R/PM (Byrne, 1998) ACT-R 5.0 (Lebiere, 2001) Windows Environment (Bothell, 2001) Macintosh Environment (Fincham, 2001) ACT-R 6.0 (Bothell, 2004??)

Other Symbolic Cognitive Architectures : 

Other Symbolic Cognitive Architectures Soar Production rule system Organized in terms of operators associated with problem spaces Goal oriented Sub-goaling Learning mechanism - Chunking EPIC -- Executive Processes in Cognition Parallel firing of production rules (in principle) Well developed visual and motor system Emphasis on executive processes

Why Model : 

Why Model Understanding and explanations Deal with complexity Span levels Explicit Make predictions Control Can look at inner workings Explore causal control

Why not? : 

Why not? Too simple Too complex Modeler must show how model uses principles to account for data Can model anything Free parameters What is the correct model Indeterminacy problem

Consciousness : 

Consciousness

Leabra Cognitive Phenomena : 

Leabra Cognitive Phenomena Parallelism Gradedness Interactivity Bidirectional connectivity Competition Selection of certain representations Inhibitory neurons Learning

ACTR Assumptions : 

ACTR Assumptions Widely parallel but serial bottleneck Two layers Symbolic layer Sub-symbolic layer Bayesian computations Declarative/Procedural Knowledge Activation based memory Utility Production Selection

Neuron : 

Neuron Basic information processing mechanisms Biological neuron - very complex Simulated neuron - simplify but extract basic functional characteristic Computational level - a detector Biological mechanisms

Detector Model : 

Detector Model Detecting some set of conditions (e.g. a smoke detector) Integrate and Fire Model Integrates inputs from different sources Into a real-valued number (its output) Reflects how well the inputs match what the neuron has become specialized to detect What does the neuron represent? What is detected Layered

Detector Model & Neural Components : 

Detector Model & Neural Components

Mapping Integrate and Fire to Detector Model : 

Mapping Integrate and Fire to Detector Model Inputs - synapses on dentrites Weights - efficiency of synaptic connection (synaptic strengths) Integration - membrane potential at the cell body Output - Threshold at start of axon

Biology of Neuron : 

Biology of Neuron Single cell Cell body, membrane and nucleus Membrane has channels Ions pass through Electric Potential (Voltage) Integrated in cell body Activates action potential output in axon Release of neurotransmitter Activates potential via dendritic synaptic input channels

The Axon : 

The Axon Spiking - action of neural firing Starts at axon hillock Refractory period - time following a spike that potential is to low to fire again Active and Passive transmission down the axon Active - domino effect, high energy, slow Passive - electrical, fast, relay stations (nodes of Ranvier)

Synapse : 

Synapse

Weight = Synaptic efficacy : 

Weight = Synaptic efficacy Synaptic efficacy = activity of presynaptic neuron communicated to postsynaptic neuron Sending: # of vesicles released, NT/vesicle, efficiency of reuptake mechanism Receiving: # of receptors, geometry of dendrite/spine, efficency of channel, proximity of release site and receptor

The Dendrite : 

The Dendrite NT - Glutamate binds to AMPA, NMDA, and mGlu receptors AMPA - excitatory - Na+ raises potential NMDA - Ca++ - learning NT - GABA - GABA receptors - Inhibitory - Cl-

Important Property : 

Important Property In the cortex, a given type of neuron releases only one type of neurotransmitter Activates particular types of receptors Therefore, cortical neurons either send excitatory inputs or inhibitory inputs but not both Can and do receive both

Neuron as an electro-chemical system : 

Neuron as an electro-chemical system Ions flow in and out due to electric and diffusion forces Membrane potential Vm - electric potential between the inside and outside of the cell Represents the integration of inputs and forces (electric and diffusion)

Basic Electricity : 

Basic Electricity Electric charge Positive and negative Opposites ?, like ? Ions have net charge Na+, Cl-, Ca++, K+ Current flows to balance positive and negative charge Voltage (potential) - difference in charge

Resistance : 

Resistance Ions encounter resistance when they move

Ohm’s Law : 

Ohm’s Law I = V / R I = current, V = voltage, and R = resistance Conductance G = 1 / R So I = VG

Diffusion : 

Diffusion Constant motion causes mixing which evens out the distribution Acts on each ion type separately

Equilibrium : 

Equilibrium Balance between electricity and diffusion so concentration stays constant E = Equilibrium potential = amount of electric potential needed to counteract diffusion: I = G(V - E) Also called Reverse potential or Driving potential

Fick’s first law : 

Fick’s first law I = - D C I = diffusion current, D = diffusion coefficient and C = concentration potential

Neuron and its Ions : 

Neuron and its Ions