Cognitive Architectures Comparision based on perceptual processing

50 %
50 %
Information about Cognitive Architectures Comparision based on perceptual processing
Education

Published on March 12, 2009

Author: sumitavamukh

Source: slideshare.net

Description

A comparision between different cognitive architectures based on how each handles perception like ACT-R, EPIC, CHREST, SOAR,ICARUS.

This is a presentation made for a talk at NAOP Annual Convention at IIT, Guwahati 2009.
(C) Sumitava Mukherjee
[smukh@cognobytes.com/ smukh@cbcs.ac.in
URL : http://people.cognobytes.com/smukh]

Perceptual Processing in Cognitive Architectures: A comparative Analysis Sumitava Mukherjee and Narayanan Srinivasan Centre of Behavioural and Cognitive Sciences, University of Allahabad www.cbcs.ac.in

Sumitava Mukherjee and Narayanan Srinivasan

Centre of Behavioural and Cognitive Sciences,

University of Allahabad

www.cbcs.ac.in

Cognitive Architectures A cognitive architecture specifies the underlying infrastructure for an intelligent system. The specification of a cognitive architecture consists of its representational assumptions, the characteristics of its memories, and the processes that operate on those memories. Research on cognitive architectures is important because it supports a central goal of artificial intelligence and cognitive science: the creation and understanding of synthetic agents that support the same capabilities as humans. Unlike expert systems, cognitive architectures aim for breadth of coverage across a diverse set of tasks and domains. More important, they offer accounts of intelligent behavior at the systems level, rather than at the level of component methods designed for specialized tasks. Def: A Cognitive architecture can be defined simply as the portion of a system that provides and manages the primitive resources of an agent.

Perception ~ What we need in Cog Arch.. We consider only visual perception is this discussion Low level fine grained processing of initial stimulus Four stages of perceptual processing – Image based, surface based, object based, category based. Depth perception Scene (Gist) analysis based on theories of distributed attention Affective states and interaction between affective states and emotion Expectations Motion (perhaps) time perception

We consider only visual perception is this discussion

Low level fine grained processing of initial stimulus

Four stages of perceptual processing – Image based, surface based, object based, category based.

Depth perception

Scene (Gist) analysis based on theories of distributed attention

Affective states and interaction between affective states and emotion

Expectations

Motion

(perhaps) time perception

Introduction to our Comparison of Cog Arch [ We have considered only Visual perception for this analysis] Factors used for Analysis: Whether initial information needs to be programmed or learned Modularity – Is perception modular or not Is the role of expectations handled Nature (granularity) of input visual representation Cognitive Architectures included in this comparative analysis: ACT-R/PM [Adaptive Control of Thought] SOAR EPIC [Executive Process/ Interactive Control ] CHREST [Chunk Hierarchy and Retrival Structures] ICARUS

[ We have considered only Visual perception for this analysis]

Factors used for Analysis:

Whether initial information needs to be programmed or learned

Modularity – Is perception modular or not

Is the role of expectations handled

Nature (granularity) of input visual representation

Cognitive Architectures included in this comparative analysis:

ACT-R/PM [Adaptive Control of Thought]

SOAR

EPIC [Executive Process/ Interactive Control ]

CHREST [Chunk Hierarchy and Retrival Structures]

ICARUS

ACT-R/PM ACT-R's main components are: modules, buffers, pattern matcher. There are two types of modules: perceptual-motor modules: takes care of the interface with the real world (i.e., with a simulation of the real world)‏ memory modules There are two kinds of memory modules: declarative memory and procedural memory Buffers: ACT-R accesses its modules (except for the procedural-memory module) through buffers. http://act-r.psy.cmu.edu/about/

SOAR Hypothesis : All deliberate goal -oriented behavior can be cast as the selection and application of operators to a state . A state is a representation of the current problem-solving situation; an operator transforms a state (makes changes to the representation); a goal is a desired outcome of the problem-solving activity. As Soar runs, it is continually trying to apply the current operator and select the next operator (a state can have only one operator at a time), until the goal has been achieved. Soar has separate memories (and different representations) for descriptions of its current situation and its long-term knowledge. In Soar, the current situation, including data from sensors, results of intermediate inferences, active goals, and active operators is held in working memory .( organized as objects). Objects are described in terms of their attributes ; the values of the attributes may correspond to sub-objects, so the description of the state can have a hierarchical organization. The Soar architecture cannot solve any problems without the addition of long-term knowledge. ( Note the distinction between the “Soar architecture” and the “Soar program”): Soar architecture” refers to the system while the “Soar program” refers to knowledge added to the architecture. Soar execution ..select-> apply..

EPIC Human performance in a task is simulated by programming the cognitive processor with production rules organized as methods for accomplishing task goals. The EPIC model then is run in interaction with a simulation of the external system and performs the same task as the human operator would. The model generates events (e.g. eye movements, key strokes, vocal utterances) whose timing is accurately predictive of human performance. Multi-task performance and its simulation in EPIC is one of the core research focus.

CHREST The model combines low-level aspects of cognition (e.g., mechanisms monitoring information in short-term memory) with high-level aspects of cognition (e.g., use of strategies) . It consists of perception facilities for interacting with the external world, short-term memory stores (in particular, visual and verbal memory stores), a long-term memory store, and associated mechanisms for problem solving. Short-term memory in CHREST contains references to chunks held in long-term memory, which are recognized through the discrimination network from information acquired by the perception system.

ICARUS The basic Icarus interpreter operates on a recognize-act cycle but, unlike many architectures, focuses on reactive execution of existing skills rather than on problem-space search. A skill consists of three elements stated in terms of logical expressions: a set of objectives, a set of requirements or preconditions, and a set of alternate means for accomplishing the objective under those conditions. Each objective, requirement, or means can refer to primitive actions/sensors or to other Icarus skills, thus imposing a hierarchical organization on long-term memory. Each skill also has an associated utility cast as a linear function of sensory attributes. Given a top-level skill to pursue, on each cycle the system first checks the objective field for that skill. If the objectives are true , nothing further needs to be done, but, if not, the interpreter examines the requirements to determine if the preconditions for action are met. If not , Icarus invokes a subskill associated with the failed requirement in an effort to satisfy it; otherwise, it selects one of the alternate means and calls on the primitive action or subskill associated with it. The architecture selects the alternative with the highest expected utility as predicted by a linear function associated with each skill.

Comparing Cog Arch

Concluding remarks Most of the architectures focus on classically higher order cognition (like learning and problem solving)‏ However, all the architectures have incorporated perception as an 'add on' but was not initially designed with the idea that cognition is rooted in perception (Except CHREST)‏ Although the architectures do have perceptual abilities, what needs to be added are: Object level perception (and depth perception)‏ Implementation of scene perception(perhaps based on distributed processing theories)‏ Emotional effects on perception (and attention)‏ Self learning mechanisms based on 'experience' Thus, although cognitive architectures are doing some domain specific tasks (mostly) quite well, lot of perceptual processing has yet to be built in.

Most of the architectures focus on classically higher order cognition (like learning and problem solving)‏

However, all the architectures have incorporated perception as an 'add on' but was not initially designed with the idea that cognition is rooted in perception (Except CHREST)‏

Although the architectures do have perceptual abilities, what needs to be added are:

Object level perception (and depth perception)‏

Implementation of scene perception(perhaps based on distributed processing theories)‏

Emotional effects on perception (and attention)‏

Self learning mechanisms based on 'experience'

Thus, although cognitive architectures are doing some domain specific tasks (mostly) quite well, lot of perceptual processing has yet to be built in.

Add a comment

Related presentations

Related pages

Cognitive architecture - Wikipedia, the free encyclopedia

Some cognitive architectures or models are based on a set of ... other low level perceptual processing ... Cognitive architecture comparison;
Read more

Including a Model of Visual Processing With a Cognitive ...

Including a model of Visual Processing with a Cognitive Architecture ... architecture of cognitive processing ... based on what cognitive, perceptual, ...
Read more

A Survey of Cognitive Architectures - CMU Computer Science

A simplified cognitive architecture for GOMS modeling. ... Perceptual Motor Cognitive Arousal ... A Survey of Cognitive Architectures 26 Summary Comparison
Read more

CiteSeerX — Citation Query Cognitive architecture

... Cognitive architecture. ... a new cognitive architecture based on recent ... from primarily cognitive processes to perceptual processing ...
Read more

Cognitive Architectures for Modeling Driver Behavior Dario ...

Cognitive Architectures for Modeling Driver Behavior ... of relevant cognitive, perceptual, ... based on cognitive architectures can be linked ...
Read more

T COGNITIVE ARCHITECTURE OF COMPLEX M - uni-bielefeld.de

Cognitive Architecture of Complex Movement Correspondence concerning this article should be sent to: ... symmetry tendency based on perceptual information.
Read more

Society of Mind” Cognitive Architecture

“Society of Mind” Cognitive Architecture ... basic CAMAL cognitive architecture with extra processing layers ... Comparison of cognitive architectures
Read more

Dr. Hakwan Lau: The Signal Processing Architectures ...

... Processing Architectures Underlying Subjective Perceptual ... comparison based on behavioral data. An indirect argument for the Hierarchical model can ...
Read more

Perceptual Representation and Reasoning - Association for ...

PERCEPTUAL REPRESENTATION AND REASONING ... comparisons with other ... operatiom that tim perceptual architecture provides on perceptual ...
Read more