adaptability theory as a guide for interfacing computers and human society

36
Adaptability Theory as a Guide for Interfacing Computers and Human Society

Upload: beverly-osborne

Post on 17-Dec-2015

222 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Adaptability Theory as a Guide for Interfacing Computers and

Human Society

Page 2: Adaptability Theory as a Guide for Interfacing Computers and Human Society

objective

• To elucidate the comparative capabilities of organisms and machines

• To clarify the potential benefits and dangers of algorithmic approaches

Page 3: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Outlines

• Hierarchical adaptability

• Adaptability

• Biological compensation principle

• Tradeoff principle

Page 4: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Hierarchical adaptability

• A tool for analyzing the mechanisms and processes that living systems use to survive in uncertain or unknown environment

Page 5: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Adaptability

• The use of information about the environment to select from a repertoire of possible behaviors

• The ability to continue function in the face of an uncertain or unknown environment

• The ability to tolerate the uncertainty of the most uncertain environment

Page 6: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Compensation principle

• Restrictions on one part of this repertoire (e.g., on genetic plasticity) must be compensated by expansions of other parts (e.g., neurobehavioral plasticity), by an increase in computing power, or by a decrease in niche space

• Restrictions on one form of adaptability must be compensated by amplification of other forms, or by naroowing of the environment

Page 7: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Tradeoff principle

• Programmability is incompatible with high evolutionary plasticity and efficient use of computational resources

• A tool for analyzing the relationship between evolutionary adaptability and information processing adaptability

Page 8: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Tradeoff principles

• Programmability: the type of control which we exert over digital computers

• Computational efficiency: the couple of the system’s resources to the solution

• Evolutionary adaptability: learn through variation and selection

• The above three parameters cannot be present at the same time in the same system

Page 9: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Compensation and Tradeoffprinciples

• Programmable computing power↑ => adaptability↓

• Programmable computing power↑ => put humans in the position of disturbance-absorbing homeostats for computers

Page 10: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Principle of requisite variety

• To cope with the uncertainty of the environment, a system must have a sufficiently large repertoire of behaviors (Ashby, 1956).

Page 11: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Biological compensation

• Simple organisms (bacteria) and higher plants: high genetic variability & simple morphology.

• Bacteria has good problem-solving & short generation times.

• Plant has less problem-solving & long generation times (adaptability↓, morphology↑ )

• Higher vertebrates: complex morphology morphology↓, specialized info. systems↑ )

Page 12: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Summary

• More complex systems => evolutionary adaptability more expensive => algorithmic problem solving increase

Page 13: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Summary

• If a system is structurally programmable, it is too sensitive to a single structural change.

• If a system is programmable, it can be used to simulate any particular system (but less efficient than the system being simulated).

Page 14: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Summary of digital computer

• Tradeoff principle: programmability

• Compensation principle: algorithm problem solving↑, evolutionary problem solving↓ , efficiency ↓

Page 15: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Adaptability

• H(W) – H(W|w)+H(w|W) H(w)≧• H(W): uncertainty of the system

• H(w):uncertainty of the environment

• H(W|w):ability of the system to anticipate the environment

• H(w|W): insensitivity of the system to the environment

Page 16: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Uncertainty of system: H(W)↑

• the repertoire of possible behaviors↑ (e.g., genetic diversity, variety of developmental and morphological patterns, immune systems, physiological plasticities, community and population plasticities

Page 17: Adaptability Theory as a Guide for Interfacing Computers and Human Society

System to anticipate environment H(W/w)

• Increase processing power of the system

• Decrease internal noises (internal uncertainty)

Page 18: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Insensitivity to environmentH(w/W)

• Decrease the niche of the environment space occupied by the system, for instance, the niche of the system

Page 19: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Adaptability rule

• H(W) – H(W|w)+H(w|W) H(w)≧• LHS: the most uncertain tolerable

environment

• RHS: the most uncertainty of environment

Page 20: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Hierarchical structures

• System: ecosystem, community, populations, organism, cells, genes, subcellular components

• Environment: environment, local regions, small compartments

Page 21: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Hierarchical environment

• H(w) = ∑He(wij)

• He(wij) = fH(H(wij))+conditional terms

• F is a normalizing coefficient

• Unconditional terms: the behavioral uncertainty of each individual subsystem

• Conditional term: the correlation between this uncertainty and other subsystems

Page 22: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Biological compensation

• Decreases in one form of adaptability must be compensated by increases in other forms of adaptability or by increases in niche breadth.

Page 23: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Adaptability and Variability

• Adaptability requires tow forms of variability.

• 1) contributes to different functional states of the system

• 2) contributes to the information processing which controls the transitions among functionally distinct states (finer form).

Page 24: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Finer form of variability

• Required continuous dynamics and redundancy

• Continuous dynamics allows gradual functional change with single genetic changes.

• Redundancy buffers the effects of genetic change on vital aspects of function.

Page 25: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Compare between nature’s and man-created technology

• Sensitivity to structural change

• Importance of gradualism for evolution

• Efficiency and complexity

Page 26: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Evolution of information system

• Human brains

• Human-created computers

Page 27: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Human brains

• Too complex to allow for wide ranging hybrid complexes and radical variations of form, too expensive to build and too long living to allow for the type of culturability that microorganisms utilized

• Complex morphological specialization↑, evolutionary adaptability↓=>amplification of special systems (nervous system)

Page 28: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Interplay between complex specializations and viable

adaptability structure• Keep brain size small as the environment

remained narrow• Enlarge brain size (costly mechanisms) to

be insensitive to environmental variations• Ecological principle of competitive

exclusion came in play: no two different species can coexist in the same niche.

• Increasing reliance on algorithmic solving

Page 29: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Human brain – shift to algorithmic solving

• Programmable at the interpretive level (read and following rules)

• Nonprogrammable at the structural level

• To analyze signals

• To conceptually and formally model

• To make predictions and plans

• anticipation↑, behavioral uncertainty↓

Page 30: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Trend

• May societies are moving rapidly to computerize their activities as much as possible.

Page 31: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Problem in current computers

Development of computing continues without any recognition of the tradeoff principle (processing power↑, adaptability↓)

Improper integration of programmable systems into society will increase rigidity at all levels of organization and decrease evolutionary mechanisms of social adaptability

Page 32: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Towards adaptable computer-based systems

• Adaptable software design

• Documentation

• Organization-guided models

• Evolutionary programming

• Molecular computers

• Neural and neuromolecular computing

• Complexity theory

Page 33: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Adaptable software design

The desire for modifiable and easily maintainable codes is a manifestation of problems emanating from the tradeoff principle.

Good programming practice helps

Page 34: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Documentation

• Setup the connection between the code and the environment in which it operates

• Adaptability is parasitic on human intelligence and and uses humans as homestatic controller for computer systems.

Page 35: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Organization-guided models

• Humans enjoy a body-guided model that each part has its specific needs and functions.

Page 36: Adaptability Theory as a Guide for Interfacing Computers and Human Society

Conclusion

• Compensation principle told us that enhanced info. processing can increase adaptability, but not in structural programming systems.

• Ill-advised shifts to structurally programmable forms of computing must either restrict the ability of society to utilize the environment, or force human beings to serve as homestatic devices for computers.