Presentation is loading. Please wait.

Presentation is loading. Please wait.

ALife Modeling of Evolution George Kampis Basler Chair, ETSU, 2007.

Similar presentations


Presentation on theme: "ALife Modeling of Evolution George Kampis Basler Chair, ETSU, 2007."— Presentation transcript:

1 ALife Modeling of Evolution George Kampis Basler Chair, ETSU, 2007

2 February 6, 2007ALife modeling of evolution2 Basler Lecture I

3 February 6, 2007ALife modeling of evolution3 Basler Lectures II-IV …, Kampis is presenting a series of free public lectures. Remaining lectures this semester will be held Feb. 6 (“The Evolution of Species in Artificial Life Models”), March 14 (“Complexity Theory in Biological and Social Systems”) and April 12 (“Consciousness in the Body”) at 7 p.m. in the Brown Hall auditorium.

4 February 6, 2007ALife modeling of evolution4 And now… For something completely different….

5 February 6, 2007ALife modeling of evolution5 The Plan The Project Artificial Life Artificial Evolution The FATINT/EvoTech system Evolution by phenotye plasticity The Model Results Future work

6 February 6, 2007ALife modeling of evolution6 The Institute

7 February 6, 2007ALife modeling of evolution7 György Kampis Group leader Professor ColBud László Gulyás Researcher (p.t.) PhD, ColBuD István Karsai Associate Director, ETSU IQB Nigel Gilbert CS Advisor Péter Érdi CS Advisor Mark Bedau CS Advisor Imre Kondor CS Advisor Katalin Mund Associate PhD Student ELTE students W. de Back PhD, ColBud Sándor Soós Coordinator PhD, ColBud

8 February 6, 2007ALife modeling of evolution8 György Kampis Group leader Professor ColBud László Gulyás Researcher (p.t.) PhD, ColBuD István Karsai Associate Director, ETSU IQB Nigel Gilbert CS Advisor Péter Érdi CS Advisor Mark Bedau CS Advisor Imre Kondor CS Advisor Katalin Mund Associate PhD Student ELTE students W. de Back PhD, ColBud Sándor Soós Coordinator PhD, ColBud

9 February 6, 2007ALife modeling of evolution9 György Kampis Group leader Professor ColBud László Gulyás Researcher (p.t.) PhD, ColBuD István Karsai Associate Director, ETSU IQB Nigel Gilbert CS Advisor Péter Érdi CS Advisor Mark Bedau CS Advisor Imre Kondor CS Advisor Katalin Mund Associate PhD Student ELTE students W. de Back PhD, ColBud Sándor Soós Coordinator PhD, ColBud

10 February 6, 2007ALife modeling of evolution10 Special thanks to..

11 February 6, 2007ALife modeling of evolution11 Artificial Life, what is it?

12 February 6, 2007ALife modeling of evolution12 Artificial Life, what is it?

13 February 6, 2007ALife modeling of evolution13 Life „in silico” J. von Neumann (1966) Self reproducing automata

14 February 6, 2007ALife modeling of evolution14 ALife modeling

15 February 6, 2007ALife modeling of evolution15 Evolution

16 February 6, 2007ALife modeling of evolution16 Evolution..is the origin of higher structure and increase of complexity

17 February 6, 2007ALife modeling of evolution17 Artificial Evolution

18 February 6, 2007ALife modeling of evolution18 Evolution in the Lab I.

19 February 6, 2007ALife modeling of evolution19 Evolution in the Lab II.

20 February 6, 2007ALife modeling of evolution20 Tierra, Avida…

21 February 6, 2007ALife modeling of evolution21 Limits from design „Running out of fuel” No emergence of higher structures No proof of complexity increase

22 February 6, 2007ALife modeling of evolution22 BioMorphs

23 February 6, 2007ALife modeling of evolution23 „Galapagos” 1997 Karl Sims

24 February 6, 2007ALife modeling of evolution24 Spore

25 February 6, 2007ALife modeling of evolution25 Spore is…

26 February 6, 2007ALife modeling of evolution26 a God game!

27 February 6, 2007ALife modeling of evolution27 Can we get any further?

28 February 6, 2007ALife modeling of evolution28

29 February 6, 2007ALife modeling of evolution29 Fun: The Evolution Prize

30 February 6, 2007ALife modeling of evolution30 The FATINT/EvoTech system

31 February 6, 2007ALife modeling of evolution31

32 February 6, 2007ALife modeling of evolution32 Evolution produces diverse species...

33 February 6, 2007ALife modeling of evolution33.. usually splitting old ones, into two...

34 February 6, 2007ALife modeling of evolution34 …or more.

35 February 6, 2007ALife modeling of evolution35 A species is a set of interbreeding individuals... Reproductively isolated from others (e.g. mechanically, behaviorally, in terms of mating preference, etc.) The key is: Species

36 February 6, 2007ALife modeling of evolution36 Can we produce new species? I.e., can we get existing species to split, A more realistic species. Not panmictic, yet stable. structurally,functionally, repeatedly, andfor internal reasons? In an evolutionary engine, a species, which responds to a selection force, just does not split.

37 February 6, 2007ALife modeling of evolution37 Simmons, Robert E. and Lue Scheepers. "Winning by a Neck: Sexual Selection in the Evolution of Giraffe." The American Naturalist Nov 96: Back to the Giraffe… Species splitting by sexual selection.

38 February 6, 2007ALife modeling of evolution38 Function change - interaction change The pre-okapi neck was always there, but was used differently function change (Darwin), exaptation (Gould), tinkering (Jacob) „niche construction” (Lewontin, Laland, Odling-Smee, Feldman……) Phenotype plasticity (Karsai, Kampis..)

39 February 6, 2007ALife modeling of evolution39 Explicit and implicit phenotype Our model: sexual selection from changing mating preference, not changing male behavior/changing use. Similar to the change of use, the change of mating preference means a change of an implicit phenotype trait into an explicit phenotype trait. This can be caused/accompanied by a mutation, or not (as in the case of phenocopies or imitative behavior change). Similar to the Baldwin effect.

40 February 6, 2007ALife modeling of evolution40 Conjecture and test Conjecture: a changing phenotype (i.e. interaction-) definition facilitiates speciation. A minimal test: study the effect of adding a new phenotype dimension.

41 February 6, 2007ALife modeling of evolution41 Finding the „right” mating partner... is a matter of fitting together. … can be modeled as template matching, Mating preference in an Alife model which defines a metric (similarity) space that can be used in any # of dimensions.

42 February 6, 2007ALife modeling of evolution42 The Model Evolution Engine Energy Interaction Change Support

43 February 6, 2007ALife modeling of evolution43 The Model Evolution Engine –agents in a non-spatial, partial artificial ecology –single resource, energy –full life-cycle (reproduction, aging, death)… –genderless sexual reproduction –standard evolutionary operators: mutation, crossing-over Interaction Change Evolution Engine Interaction Change Energy

44 February 6, 2007ALife modeling of evolution44 The Model Evolution Engine –agents in a non-spatial, partial artificial ecology –single resource, energy –full life-cycle (reproduction, aging, death)… –genderless sexual reproduction –standard evolutionary operators: mutation, crossing-over Interaction Change Uptake Consumption Reproduction Death

45 February 6, 2007ALife modeling of evolution45 The Model Evolution Engine –agents in a non-spatial, partial artificial ecology –single resource, energy –full life-cycle (reproduction, aging, death)… –genderless sexual reproduction –standard evolutionary operators: mutation, crossing-over Interaction Change

46 February 6, 2007ALife modeling of evolution46 The Model Evolution Engine –agents in a non-spatial, partial artificial ecology –single resource, energy –full life-cycle (reproduction, aging, death)… –genderless sexual reproduction –standard evolutionary operators: mutation, crossing-over Interaction Change

47 February 6, 2007ALife modeling of evolution47 Evolution Engine Agent: –Phenotype vector in N-dimensional space

48 February 6, 2007ALife modeling of evolution48 Interaction Change Agent: –Phenotype vector in N+1-dimensional space. Type-independent, and Type-based methods

49 February 6, 2007ALife modeling of evolution49 Emergent Phenotypes Form Cause Type Point mutationendog.local Phenocopiesexog.part global Epigenetic changebothpart global Horizontal adapt.bothglobal Behavior changesocialglobal in natural and in model populations

50 February 6, 2007ALife modeling of evolution50 Results

51 February 6, 2007ALife modeling of evolution51 Species splitting

52 February 6, 2007ALife modeling of evolution52

53 February 6, 2007ALife modeling of evolution53 Ordinations

54 February 6, 2007ALife modeling of evolution54 The production of species type-dependent method type-independent method number of dimensions

55 February 6, 2007ALife modeling of evolution55 Conclusions The model yields simulation results in an interactions- based evolutionary model, which uses similarity-based sexual selection to approach open ended evolution. It demonstrates that changing phenotype-to-phenotype interaction (here: the adding of new dimensions to interaction space) can repeatedly split species by producing new selection constraints. We hypothesize that sustained ecological evolution proceeds by similar mechanisms.

56 February 6, 2007ALife modeling of evolution56 Future work Developmental modeling Ecosystems 3D physics

57 February 6, 2007ALife modeling of evolution57 Aplysia

58 February 6, 2007ALife modeling of evolution58

59 February 6, 2007ALife modeling of evolution59

60 February 6, 2007ALife modeling of evolution60

61 February 6, 2007ALife modeling of evolution61 Dis/similarity in ecosystems.

62 February 6, 2007ALife modeling of evolution62 Can relational properties bootstrap whole ecosystems?

63 February 6, 2007ALife modeling of evolution63 3D physics Using


Download ppt "ALife Modeling of Evolution George Kampis Basler Chair, ETSU, 2007."

Similar presentations


Ads by Google