Difference between revisions of "ArchitectureModelV2"

From aHuman Wiki
Jump to: navigation, search
(Automated page entry using MWPush.pl)
 
(Automated page entry using MWPush.pl)
 
(4 intermediate revisions by the same user not shown)
Line 1: Line 1:
−
 
 
<pre style="color: green">High Level Architectural Design version 2</pre>
 
<pre style="color: green">High Level Architectural Design version 2</pre>
  
 
@@[[Home]] -> [[Architecture]] -> [[ArchitectureModelV2]]
 
@@[[Home]] -> [[Architecture]] -> [[ArchitectureModelV2]]
  
−
<wiki:toc max_depth="3" />
+
__TOC__
 
----
 
----
  
Line 15: Line 14:
 
== Mapping To Human Body ==
 
== Mapping To Human Body ==
  
−
http://ahuman.googlecode.com/svn/images/wiki/development/humanmind.jpg
+
http://usvn.ahuman.org/svn/ahwiki/images/wiki/development/humanmind.jpg
  
 
== Brain Components ==
 
== Brain Components ==
Line 21: Line 20:
 
Informational model of real brain (actually left body is connected to right brain - but it does not bring any difference from model perspective):
 
Informational model of real brain (actually left body is connected to right brain - but it does not bring any difference from model perspective):
  
−
http://ahuman.googlecode.com/svn/images/wiki/research/brain/RealBrain.jpg
+
http://usvn.ahuman.org/svn/ahwiki/images/wiki/research/brain/RealBrain.jpg
  
 
Below aggregated model demonstrates spinal cord role and principal mind flows.
 
Below aggregated model demonstrates spinal cord role and principal mind flows.
Line 27: Line 26:
 
Left and right items are shown as double blocks.
 
Left and right items are shown as double blocks.
  
−
http://ahuman.googlecode.com/svn/images/wiki/research/brain/AggregatedBrain.jpg
+
http://usvn.ahuman.org/svn/ahwiki/images/wiki/research/brain/AggregatedBrain.jpg
  
 
Model reflects cognitive, conscious and biological processes.
 
Model reflects cognitive, conscious and biological processes.
Line 33: Line 32:
 
Find more detailed informational model of components which implement cognition:
 
Find more detailed informational model of components which implement cognition:
  
−
http://ahuman.googlecode.com/svn/images/wiki/research/brain/CortexLinks.jpg
+
http://usvn.ahuman.org/svn/ahwiki/images/wiki/research/brain/CortexLinks.jpg
  
 
== Current understanding ==
 
== Current understanding ==
Line 47: Line 46:
 
* Items should demonstrate ability to adopt to environment "by design"
 
* Items should demonstrate ability to adopt to environment "by design"
  
−
http://ahuman.googlecode.com/svn/images/wiki/development/mindmap.jpg
+
http://usvn.ahuman.org/svn/ahwiki/images/wiki/development/mindmap.jpg
  
 
== Function / Technology Map ==
 
== Function / Technology Map ==
  
−
|| *Module / Component Set* || *Primary Function* || *Related Problems and Algorithms* ||
+
{| class=wikitable width=80%
−
|| *engine* || startup sequence and component management || ||
+
| '''Module / Component Set''' || '''Primary Function''' || '''Related Problems and Algorithms'''
−
|| *modbody* || || ||
+
|-
−
|| physiology || positioning of sensors and effectors || ||
+
| '''engine''' || startup sequence and component management ||
−
|| biology || internal senses and motivation || ||
+
|-
−
|| sensors || produce raw sensor flow || chat reader, video-stream, signal processing ||
+
| '''modbody''' || ||
−
|| effectors || pre-defined sets of actions in environment || chat writer, primary feedback ||
+
|-
−
|| *modbrain* || multi-neurex map and its evolution || ||
+
| physiology || positioning of sensors and effectors ||
−
|| *modmedia* || physical access to external environment || ||
+
|-
−
|| *modcognition* || || ||
+
| biology || internal senses and motivation ||
−
|| neocortex || perception of sensor data || ||
+
|-
−
|| associations || associative memory || inferior temporal cortex (ITC) ||
+
| sensors || produce raw sensor flow || chat reader, video-stream, signal processing
−
|| *modintelligence* || || ||
+
|-
−
|| behavior || rational intentions from perception || ||
+
| effectors || pre-defined sets of actions in environment || chat writer, primary feedback
−
|| motor || effectors control || ||
+
|-
−
|| *svcdb* || persistence services || ||
+
| '''modbrain''' || multi-neurex map and its evolution ||
−
|| *svcio* || internal communications || JMS/COM-style ||
+
|-
−
|| *lib* || || ||
+
| '''modmedia''' || physical access to external environment ||
−
|| libgen || low-level generic algorithms and data primitives || ||
+
|-
−
|| libnn || neural networks of various types || HMM, ANN ||
+
| '''modcognition''' || ||
−
|| libbn || belief networks || HTM ||
+
|-
 +
| neocortex || perception of sensor data ||
 +
|-
 +
| associations || associative memory || inferior temporal cortex (ITC)
 +
|-
 +
| '''modintelligence''' || ||
 +
|-
 +
| behavior || rational intentions from perception ||
 +
|-
 +
| motor || effectors control ||
 +
|-
 +
| '''svcdb''' || persistence services ||
 +
|-
 +
| '''svcio''' || internal communications || JMS/COM-style
 +
|-
 +
| '''lib''' || ||
 +
|-
 +
| libgen || low-level generic algorithms and data primitives ||
 +
|-
 +
| libnn || neural networks of various types || HMM, ANN
 +
|-
 +
| libbn || belief networks || HTM
 +
|}
  
−
*Major Diagram Features*:
+
'''Major Diagram Features''':
  
 
* Diagram is the same for biological human and for artificial life
 
* Diagram is the same for biological human and for artificial life
 
* Body is regarded as consisting of non-overlapping brain, physiology and biology parts
 
* Body is regarded as consisting of non-overlapping brain, physiology and biology parts
−
  * *physiology* - physiological part is not having any information processing, but serve for direct converting neural signals to musculus changes or transform light into eye signals (not including any pattern processing, capture only); in aHuman it can be created as hard-coded, non-adoptable software without loss of Artificial Life features
+
*** '''physiology''' - physiological part is not having any information processing, but serve for direct converting neural signals to musculus changes or transform light into eye signals (not including any pattern processing, capture only); in aHuman it can be created as hard-coded, non-adoptable software without loss of Artificial Life features
−
  * *biology* - non-brain biological part represents human component and subsystems, which implement human as biological being - reproductive system, feed subsystem, blood subsystem and others; in aHuman it has meaning only for functions that link brain and body - like metabolism, pain feedback
+
*** '''biology''' - non-brain biological part represents human component and subsystems, which implement human as biological being - reproductive system, feed subsystem, blood subsystem and others; in aHuman it has meaning only for functions that link brain and body - like metabolism, pain feedback
−
  * *brain* is a biological part intended to support mind operations; in aHuman it is represented by computer science solutions and algorithms based primarily on connectivists principles  
+
*** '''brain''' is a biological part intended to support mind operations; in aHuman it is represented by computer science solutions and algorithms based primarily on connectivists principles  
 
* Mind on brain
 
* Mind on brain
−
* mind is shown as cognition and intelligence functions, implemented on brain
+
** mind is shown as cognition and intelligence functions, implemented on brain
−
* several functions were identified by software model of brain
+
** several functions were identified by software model of brain

Latest revision as of 18:51, 28 November 2018

High Level Architectural Design version 2

@@Home -> Architecture -> ArchitectureModelV2

Up to the moment several models were created, addressing different sides. Latest models include more enhanced understanding of subject area. I regard ENTHORINAL cortex separate from NEOCORTEX, though from neurobiology point of view, it is neocortex as well. It is done to stress its close relation to HIPPOCAMPUS. Jeff Hawkins, in turn, regards neurobiology neocortex, THALAMUS and HIPPOCAMPUS as sufficient set of elements to explain functions of human intelligence. All brain parts, except HYPOTHALAMUS and CEREBELLUM, exist in both left and right hemispheres. Purely biological parts, like Pineal Gland are not present in model, even in aggregated form.

Mapping To Human Body

humanmind.jpg

Brain Components

Informational model of real brain (actually left body is connected to right brain - but it does not bring any difference from model perspective):

RealBrain.jpg

Below aggregated model demonstrates spinal cord role and principal mind flows. NEOCORTEX from above model is split into CT/CORTEX and FC/FRONTAL CORTEX. Left and right items are shown as double blocks.

AggregatedBrain.jpg

Model reflects cognitive, conscious and biological processes. Cognition is mostly related to neocortex, including recognition and learning. Find more detailed informational model of components which implement cognition:

CortexLinks.jpg

Current understanding

Last findings:

Now the role and interrelationships of major scientific areas - Artificial Life, Artificial Intelligence, Cognition, Neurobiology - are quite clear and the design is based on:

  • Plausible compared to human - local connections, major components are based on human design
  • Design should allow using ordinary computer, not Blue Brain one
  • All design items on first level should be based on limitations and existing human design - no attempt to descend to certain final component model
  • Design should take into account priority of life features over intelligence features
  • Items should demonstrate ability to adopt to environment "by design"

mindmap.jpg

Function / Technology Map

Module / Component Set Primary Function Related Problems and Algorithms
engine startup sequence and component management
modbody
physiology positioning of sensors and effectors
biology internal senses and motivation
sensors produce raw sensor flow chat reader, video-stream, signal processing
effectors pre-defined sets of actions in environment chat writer, primary feedback
modbrain multi-neurex map and its evolution
modmedia physical access to external environment
modcognition
neocortex perception of sensor data
associations associative memory inferior temporal cortex (ITC)
modintelligence
behavior rational intentions from perception
motor effectors control
svcdb persistence services
svcio internal communications JMS/COM-style
lib
libgen low-level generic algorithms and data primitives
libnn neural networks of various types HMM, ANN
libbn belief networks HTM

Major Diagram Features:

  • Diagram is the same for biological human and for artificial life
  • Body is regarded as consisting of non-overlapping brain, physiology and biology parts
      • physiology - physiological part is not having any information processing, but serve for direct converting neural signals to musculus changes or transform light into eye signals (not including any pattern processing, capture only); in aHuman it can be created as hard-coded, non-adoptable software without loss of Artificial Life features
      • biology - non-brain biological part represents human component and subsystems, which implement human as biological being - reproductive system, feed subsystem, blood subsystem and others; in aHuman it has meaning only for functions that link brain and body - like metabolism, pain feedback
      • brain is a biological part intended to support mind operations; in aHuman it is represented by computer science solutions and algorithms based primarily on connectivists principles
  • Mind on brain
    • mind is shown as cognition and intelligence functions, implemented on brain
    • several functions were identified by software model of brain
Retrieved from "http://wiki.ahuman.org/index.php?title=ArchitectureModelV2&oldid=11741"