ICW - H7
72 flashcards · Other · by Jaro Lubberding
Connectionism
The study of mental operations through the construction and testing of artificial neural networks.
Artificial Neural Network (ANN)
A computer simulation of how populations of real neurons might perform some task.
Serial processors
A type of information processing where one computation is performed at a time. The results of one computation typically serve as the input for the next.
Parallel Distributed Processing (PDP)
A type of information processing where large numbers of computing units perform their calculations simultaneously. Computational units can receive and process multiple inputs and transmit multiple outputs.
Knowledge-based Approach
A type of problem-solving method in which one conceptualizes the problem and its solution in terms of symbols and transformations on the symbols; representative of the traditional or classical approach in cognitive science.
Behavior-based Approach
A type of problem-solving method in which a network is allowed to produce a solution on its own. This does not involve the use of symbols. It is representative of the connectionist approach in cognitive science.
Distributed Representation
A coding in which an object or feature is represented by a pattern of activation among a group of cells or nodes.
Local Representation
Representation in the form of activation or activity in a single node in a network.
Node
A representation of a concept or element of the world.
Links
The connection between each node in an artifical network.
Weights
Values on links that determine the strenght of the connection between the nodes. They can range in value from -1.0 to +1.0.
Cell Assembly
According to Hebb, a small group of neurons that repeatedly stimulate themselves.
Phase Sequence
According to Hebb, a group of connected cell assemblies that fire together or close in time.
Perceptron
Artificial neural networks that detect and recognize information about the world, store this information, and use it in some fashion. They are characterized by the ability to learn from experience and can alter their connection strengths.
The teacher
A corrective mechanism that compares actual with desired output and generates an error signal.
Hidden layer
The second layer of a three-layer network. This is where the input layer sends its signals. It performs intermediary processing.
Error signal
In artificial neural networks, the difference between the actual and the observed output. The error signal is used to alter connection strengths and train the network.
Generalized delta rule / Back-propagation
A way of training a network by repeatedly applying error feedback to alter connection strengths. The feedback is the difference between the actual and the desired output.
Biological plausibility
The idea that artificial neural networks effectively represent and model the characteristics of real-world brains.
Graceful degradation
A slow and gradual decrease in the performance of a network with increased damage.
Interference
The inability of a network to distinguish similar patterns from one another.
Generalization
The ability to apply a learned rule to a novel situation.
Convergent Dynamics
The state of a network represented by how the values of it's weights change over time.
Stability-Plasticity Dilemma
A scenario where a network should be plastic to store novel input patters. At the same time, it should be stable to protect previously encoded patterns from being erased.
Catastrophic interference
This occurs when a network has learned to recognize an old set of patterns and then is called on to learn a new set. The learning of the new set modifies the weights of the network in such a way that the old set is forgotten.
Supervised networks
These are networks that are presented with target answers for each pattern they receive as input. The network 'knows' what the right answer is on each training trial.
Semantic networks
A network where each node has a specific meaning and, therefore, employs local representation of concepts.
Spreading activation
The activity that spread outward from nodes along links to activate other nodes.
Retrieval cues
A phenomenon in which an item related to one that was memorized can lead to successful recall.
Priming
Occurs when processing of a stimulus is facilitated by the prior representation of a related stimulus.
Hierarchical organization
A semantic network where concept nodes are arranged in different levels along a continuum from abstract to concrete.
Sentence verification
A procedure in which participants judge the truth or falsity of sentences by pushing one of two buttons.
Superordinate
The most abstract form of conceptual category organization that encompasses all examples of the concept.
Ordinate
A level of concept category organization of moderate specificity.
Subordinate
The most concrete or specific form of conceptual category organization.
Prototypes
A generic or idealized representation of a conceptual category.
Cognitive economy
The principle that concepts should not have to be coded for more times than is necessary.
"Isa" link
A link in a propositional network that represents relationships of belonging.
"Hasa" link
A link in a propositional network that represents property relationships.
Hybrid model
A robotic architecture that employs a combination of reactive behaviors and planning or other cognitive components.
Agent link
A link in a propositional network that specifies the subject of the sentence, the one performing some action.
Object link
A link in a propositional network that denotes the object or thing to which an action is directed.
Relation link
A link in a propositional network specifying the type of relation between agent and object.
Type
Referring to an entire category.
Token
Referring to a specific instance of a category; an identification symbol that is useful for symbolic manipulation of knowledge.
T.O.T. Phenomenon
This is an acronym for "Tip of the Tongue", meaning one feels familiarity with an item but cannot quite recall it.
Degree of fan
A concept that is related to many others has a high degree of fan. The node representing the concept will subsequently have many links radiating outward from it to other nodes.
Guided search
A forceful or willed search for a memory item based on intelligence and reasoning.
Reconstructive memory
When recall of an item from memory is based on guided search and subject to bias by subsequent information.
Network science
The scientific interdisciplinary study of how complex networks operate.
Centrality
How a network coordinates information. This can be accomplished through a "leader" that receives information, evaluates it and issues commands.
Hierarchical networks
Organized into different levels. The number of nodes decreased traveling upward with each group of nodes converging onto another group above it with fewer elements.
Simple cells
Cells in the primary visual cortex that code for oriented line segments.
Complex cells
Cells in the visual system that code for an oriented line segment moving in a particular direction.
Hypercomplex cells
Cells in the visual system that code for angles (two conjoined oriented line segments) moving in a particular direction.
Small-world network
Any network where the number of links between any pair of nodes is few.
Random networks
Networks where the connections are entirely local and can, therefore, be both sort and long distance.
Ordered network
All the connections are local. Nodes are connected only to other nodes in their immediate neighborhood.
Egalitarian networks
A type of small-world network that contains mostly local short-distance connections with a few random long-distance connections.
Artistocratic networks
A type of small-world network where long-distance connections between nodes are the result of hubs.
Hubs
Nodes that have many more links than the rest of the network's nodes.
Preferential attachment
The principle by which nodes turn into hubs. Those with more links will gradually be linked to more often because they serve as good portals to gain acces to large numbers of other nodes.
Plasticity
The ability of the brain to recover and reorganize following injury, primarily by rerouting or regrowth of damaged pathways.
Clustering
The probability that two nodes linked to a third will also be linked to each other. This network statistic reflects the interconnectedness of neighborhoods.
Synchrony
Coordinated patterns of network activity. The firing rate of distant nodes can synchronize, as can more complicated frequency and timing relationships.
Percolation
The propagation of a signal through a network. For example, the way in which a disease can travel through a social network, infecting other individuals.
Percolating cluster
A single giant group of susceptible sites connected by open bonds that occupies the majority of a network.
Seizure
An uncontrolled surge of synchronous electrical activity in the brain.
Disorganized thinking
A symptom of schizophrenia characterized by a loose association between ideas. It can manifest as "word salad" in which concepts are not clearly related to each other.
Divergent thinking
A style of thinking where the goal is to generate many different ideas about a topic in a short period of time.
Free association
What happens when one is presented with an idea and asked to produce whatever related ideas come to mind without censoring or inhibiting.
Affect-infusion model
A model developed to explain the conditions under which emotions and moods infuse (influence) cognitive processing.