Jul 11, 2008 - Early exemplar: Karl Sims' Evolving Virtual Creatures (1991-4) .... 3) Sensors (how aspects of the environment (and the organism itself) are.
The EvoGrid
An Evolution Grid in Cyberspace Grey Thumb/British Computer Society, July 11, 2008 Bruce Damer
is Biota? ´ What is Artificial Life, and where is A-Life today? ´ What is the EvoGrid? ´ How would the EvoGrid affect the way we think about ourselves and our place in the universe? ´ What
Biota.org (version 1.0) A multi-disciplinary visionary conference series 1997-2001 paleontology, artificial life, simulation, virtual worlds, art, game design, science fiction
1997 – Banff Canada, Burgess Shale
1998 –Cambridge UK
1999 – San Jose CA
2001 – Berkeley CA
1997: Digital Biota Conferences: Digital Burgess (Banff Centre)
1997: Digital Biota Conferences: Digital Burgess (Banff Centre)
1998: Digital Biota 2 (Cambridge UK, Magdalene College)
1999: Digital Biota 3 (San Jose State University)
2001: Digital Biota 4 (UC Berkeley, NA Paleontological Convention)
Biota.org (version 2.0) Biota Podcast by Tom Barbalet (History, news, discussion about artificial life)
What is Artificial Life And what is the status of A-Life today?
Early exemplar: Karl Sims’ Evolving Virtual Creatures (1991-4)
Artificial Life: Exemplar: Karl Sims’ Evolving Virtual Creatures
Creatures with genomes evolving in the simulated physics of a Connection Machine.
Artificial Life: More recent projects: Noble Ape
By Tom Barbalet, a number of autonomous simulation components including a landscape simulation, biological simulation, weather simulation, sentient creature (Noble Ape) simulation and a simple intelligent-agent scripting language (ApeScript).
Artificial Life: Grey Thumb A-life “clubs” growing, Grey Thumb Boston
Artificial Life: More recent projects: Grey Thumb Boston-Nanopond
A simple hypercycle that has evolved within a modified variant of the Nanopond evolvable instruction set virtual machine called Nanopond-MV. By Adam Ierymenko.
Artificial Life: More recent projects: Grey Thumb Boston-Ant farm colony
Ants collect aphids and food in an ant colony simulation written by Brian Peltonen.
Artificial Life: More recent projects: Grey Thumb Boston-Robot Vision System
Robot vision system using evolved algorithms annotating an image for depth and boundaries. By Martin C. Martin.
Artificial Life: More recent projects: Grey Thumb Boston-Breve Evolved Agents
A physically simulated evolved agent in breve learns to walk. By Jonathan Klein.
Artificial Life: More recent projects: Darwin’s Park, University of Paris
“Growing” L-System Garden with photo and chemo-tropism.
Artificial Life: Gerald de Jong: Darwin@Home
Evolving virtual creatures a la Karl Sims, using elastic interval geometry (tensegrity structures).
Artificial Life: Gerald de Jong: Darwin@Home
Artificial Life: Will Wright - Spore
Artificial Life… NOT!
EvoGrid: A new initiative for the Artificial Life community Ending the isolations of the Walled Gardens
Early artificial life Grids: 199194, Karl Sims evolving virtual creatures on Connection Machine (2K processors), and Tom Ray’s Tierra, running across the Internet on servers (1992-98). World of Warcraft, Second Life today are all grids.
What would an artificial life Grid for the 21st Century look like? Running across the modern Internet: XML semantic spaces, web 2.0 interfaces
EvoGrid: In Two Flavours Broad EvoGrid Connecting existing A-life simulations to observe emergent behavior Deep EvoGrid Hoyle/Gordon’s “Origin of Artificial Life” simulation. Starting from “the void” and enabling a “cellular” structure and copying/mutation mechanism to emerge spontaneously.
Broad EvoGrid: a new initiative for the Artificial Life community Concept development stage (Q1-Q2 2008) Imagine an L-System forest, a herbivore simulation and a carnivore simulation all developed separately without each having its own graphical front end. Each object in the separate simulations would communicate locally or via the network using some agreed upon protocol. Next, picture one or more 3D front end “view portals” with all the bells & whistles that visualize what is going on in the engines and traffic, putting any local “area” together into a coherent scene. If it existed, such an A-life system could be run as a true grid, an “Evolution Grid” or “EvoGrid” which would take advantage of: - Free from the tyranny of the render cycle clock - Use multi-core processors - Multiple engines, scenegraphs - Projects don’t go extinct once they become citizens of the grid - Whole system complexity and adaptation grows faster than individual parts
Broad EvoGrid: Concept Development – discussants so far: Tom Barbalet, Gerald de Jong, Jeffrey Ventrella, Robert Rice, Bruce Damer. Inviting more participants from Grey Thumb and beyond.
What are the distributable atomic components of an EvoGrid? 1) Physics (laws determining how objects and energies change and interact over space and time) 2) Genotype (determines 3, 4, 5, and 6 below) 3) Sensors (how aspects of the environment (and the organism itself) are perceived and fed to the brain) 4) Brain (takes sensor data, process it, and then affects the actuators) 5) Actuators (what the brain affects) 6) Geometry (organism bodies (objects) consisting of 3D coordinates, polygons, and parametric primitives (if any).) 7) Rendering (It's sole job should be to render the geometry)
Broad EvoGrid: Concept Development – First cut XML Implementation Sample EvoGrid landscape and creature definitions (ref Tom Barbalet) female 78.59 1.01 awake moving
Deep EvoGrid: Concept Development The Problem: All existing A-life systems code in a great deal of our own “artifice”: suppositions, biases and simplistic structural understanding about our conceptions of what a living system is. We “seed” and then sometimes “tweak” A-life systems and therefore they become suspect as truly representative of biology or evolution. A biologist might therefore rightly claim that they do not tell us much about life but more about the imaginations of clever programmers and their skill at devising culling or optimization algorithms.
Deep EvoGrid: Concept Development The Challenge: A profoundly more interesting simulation might be embodied by a “Deep” EvoGrid in which we would set up a virtual sand-box, particle field or primordial soup with a physics that might mimic the real world of wet chemistry. Thereafter the simulation would be run, hands-off (no Artificial Gods please), and a “detector” set up to determine when something interesting has happened. The “toy laws of nature” could be varied for the entire simulation only at the beginning of simulation runs (rules of order).
Deep EvoGrid: Concept Development After the simulation (grid) is off and running one what would one look for? A replicator using a blueprint to construct, maintain and duplicate a container of the works?
Deep EvoGrid: Concept Development Hoyle/Gordon’s Challenge: “Order does not spontaneously form from disorder. A tornado passing through a junkyard would never assemble a 747.” Hoyle, Fred, 1983. The Intelligent Universe. New York: Holt, Rinehart, pp. 18-19.
However we are not seeking to assemble a 747. If we could merely show that the spontaneous emergence of something akin to an abstraction of a cell could occur, would this be of value to biology, philosophy, or cosmology? Richard Gordon: engage in “The Origin of Artificial Life”
Deep EvoGrid: Implementation (?)
EvoGrid: Philosophical Implications Will biologists (one day) declare these environments “worthy of study”? Would an EvoGrid and harnessing the power of evolution become a tool for Humanity in the 21st Century? Would it become a mechanism for life’s expansion into the Solar System or for the survival and extension of life on Earth? How does a successful origin of life simulation affect our belief/not belief of God, our place in the Universe and the future of life?
EvoGrid: Project Next Steps? Framing the history, design issues, implementation possibilities, intellectual and philosophical issues around an EvoGrid. Engaging a diverse community in the EvoGrid (Broad and Deep). Prototyping and EvoGrid. Presenting the EvoGrid in the public sphere. Encouraging or initiating the building of an EvoGrid.
EvoGrid Applications? 2007: Design for a human mission to an asteroid
EvoGrid Applications? Evolved robotics for low gravity exploration
This Effort is Dedicated to Douglas Adams Biota 2: “Is there an Artificial God” The computer forms a third age of perspective, because suddenly it enables us to see how life works. Now that is an extraordinarily important point because it becomes self-evident that life, that all forms of complexity, do not flow downwards, they flow upwards and there's a whole grammar that anybody who is used to using computers is now familiar with, which means that evolution is no longer a particular thing, because anybody who's ever looked at the way a computer program works, knows that very, very simple iterative pieces of code, each line of which is tremendously straightforward, give rise to enormously complex phenomena in a computer - and by enormously complex phenomena, I mean a word processing program just as much as I mean Tierra or Creatures.
Broad EvoGrid: Concept Development – Goals
How to get to Broad EvoGrid Alpha 0.1? 1) Design Taxonomy, Semantics and Protocols (simple tags, minimal subset) 2) Connect a couple of existing Alife simulations into the Grid together with a Visualizer: How about active Grey Thumb projects, Noble Ape, Darwin@Home, Darwin’s Pond, Darwin’s Park, Nerves Router/Digital Spaces or Breve? 4) Build a simple demonstration grid by end of 2008?
Interested? Sign up on mailing list, blog, and Biota Podcasts/LIVE shows about the EvoGrid (already underway). Visit www.evogrid.org
