FROM:http://en.wikipedia.org/wiki/Hugo_de_Garis
Hugo de Garis From Wikipedia, the free encyclopedia Jump to: navigation, search
Hugo de Garis (born 1947, Sydney, Australia) is a researcher in the sub-field of artificial intelligence (AI) known as evolvable hardware. He became known in the 1990s for his research on the use of genetic algorithms to evolve neural networks using three dimensional cellular automata inside field programmable gate arrays.He claimed that this approach would enable the creation of what heterms "artificial brains" which would quickly surpass human levels ofintelligence.[1]
He has more recently been noted for his belief that a major warbetween the supporters and opponents of intelligent machines, resultingin billions of deaths, is almost inevitable before the end of the 21stcentury.[2]:234 He suggests AIs may simply eliminate the human race, and humans would be powerless to stop them because of technological singularity.This prediction has attracted debate and criticism from the AI researchcommunity, and some of its more notable members, such as Kevin Warwick, Bill Joy, Ken MacLeod, Ray Kurzweil, Hans Moravec, and Roger Penrose, have voiced their opinions on whether or not this future is likely.
de Garis originally studied theoretical physics, but he abandoned this field in favour of artificial intelligence. In 1992 he received his PhD from Université Libre de Bruxelles, Belgium. He worked as a researcher at Advanced Telecommunications Research institute international (ATR), Japan from 1994-2000, a researcher at Starlab, Brussels from 2000-2001, and associate professor of computer science at Utah State University from 2001-2006. He is currently a professor at Xiamen University, where he teaches theoretical physics and computer science, and runs the Artificial Brain lab.
Evolvable hardwareFrom 1993 to 2000 de Garis participated in a research project atATR's Human Information Processing Research Laboratories (ATR-HIP)which aimed to create a billion neuron artificial brain by the year2001.[1]The project was known as "cellular automata machine brain", or"CAM-Brain". During this 8 year span he and his fellow researcherspublished a series of papers in which they discussed the use of genetic algorithms to evolve neural structures inside 3D cellular automata.They argued that existing neural models had failed to produceintelligent behaviour because they were too small, and that in order tocreate "artificial brains" it was necessary to manually assemble tensof thousands of evolved neural modules together, with the billionneuron "CAM-Brain" requiring around 10 million modules;[3] this idea was rejected by Igor Aleksander, who said "The point is that these puzzles are not puzzles because our neural models are not large enough".[4]
Though it was initially envisaged that these cellular automata wouldrun on special computers, such as MIT's "Cellular Automata Machine-8"(CAM-8), by 1996 it was realised that the model originally proposed,which required cellular automata with thousands of states, was toocomplex to be realised in hardware. The design was considerablysimplified, and in 1997 the "collect and distribute 1 bit"("CoDi-1Bit") model was published, and work began on a hardwareimplementation using Xilinx XC6264 FPGAs. This was to be known as the "CAM Brain Machine" (CBM).[5]
The researchers evolved cellular automata for several tasks (using software simulation, not hardware):[5]
- Reproducing the XOR function.
- Generating a bitstream that alternates between 0 and 1 three times (ie. 000..111..000..).
- Generated a bitstream where the output alternates, but can bechanged from a majority of 1s to a majority of 0s by toggling an input.
- Discriminating between two square wave inputs with a different period.
- Discriminating between horizontal lines (input on a 2D grid) and random noise.
Ultimately the project failed to produce a functional robot controlsystem, and ATR terminated it along with the closure of ATR-HIP inFebruary 2001.[6]
The original aim of de Garis' work was to establish the field of"brain building" (a term of his invention) and to "create a trilliondollar industry within 20 years". Throughout the 90s his papers claimedthat by 2001 the ATR "Robokoneko" (translation: kitten robot) projectwould develop a billion-neuron "cellular automata machine brain"(CAM-brain), with "computational power equivalent to 10,000 pentiums"that could simulate the brain of a real cat. de Garis received a US$0.4million "fat brain building grant" to develop this[7].The first "CAM-brain" was delivered to ATR in 1999. After receiving afurther US$1 million grant at Starlab de Garis failed to deliver aworking "brain" before Starlab's bankruptcy. At USU de Garis announcedhe was establishing a "brain builder" group to create a secondgeneration "CAM-brain".
Current researchde Garis published his last "CAM-Brain" research paper in 2002.[8]He still works on evolvable hardware. Using a Celoxica FPGA board hesays he can create up to 50,000 neural network modules for less than$3000.
Since 2002 he has co-authored several papers on evolutionary algorithms.
He believes that topological quantum computing is about to revolutionize computer science, and hopes that his teaching will help his students to understand its principles.[9]
In 2008 de Garis received a 3 million Chinese yuan grant (around $436,000) to build an artificial brain for China (the China-Brain Project), as part of the Brain Builder Group at Wuhan University.[10]
Employment historyde Garis' original work on "CAM-brain" machines was part of an 8year research project, from 1993 to 2000, at the ATR Human InformationProcessing Research Laboratories (ATR-HIP) in Kyoto Prefecture, Japan.de Garis left in 2000, and ATR-HIP was closed on 28 February 2001. deGaris then moved to Starlabin Brussels, where he received a million dollars in funding from thegovernment of Belgium ("over a third of the Brussels government's totalbudget for scientific research", according to de Garis).[11]Starlab went bankrupt in June 2001. A few months later de Garis wasemployed as an associate professor at the computer science departmentof Utah State University. In May 2006 he became a professor at Wuhan University's international school of software, teaching graduate level pure mathematics, theoretical physics and computer science.
Since June 2006 he has been a member of the advisory board of Novamente, a commercial company which aims to create strong AI.
Cosmists and Terransde Garis believes that a major war before the end of the 21st century, resulting in billions of deaths, is almost inevitable.[2]:234Intelligent machines (or 'artilects', a shortened form of 'artificialintellects') will be far more intelligent than humans and will threatento attain world domination,resulting in a conflict between 'Cosmists', who support the artilects,and 'Terrans', who oppose them (both of these are terms of hisinvention). He describes this conflict as a 'gigadeath' war,reinforcing the point that billions of people will be killed.[12]This scenario has been criticised by other AI researchers, includingChris Malcolm, who described it as "entertaining science fiction horrorstories which happen to have caught the attention of the popular media".[13] Kevin Warwick called it a "hellish nightmare, as portrayed in films such as the Terminator".[2]:back cover In 2005 de Garis published a book describing his views on this topic entitled The Artilect War.[2]
Cosmism is a moral philosophy that favours building or growing strong artificial intelligenceand ultimately leaving the planet Earth to the Terrans, who oppose thispath for humanity. The first half of the book describe technologieswhich he believes will make it possible for computers to be billions ortrillions of times more intelligent than humans. He predicts that asartificial intelligence improves and becomes progressively morehuman-like, differing views will begin to emerge regarding how far suchresearch should be allowed to proceed. Cosmists will foresee themassive, truly astronomical potential of substrate-independentcognition, and will therefore advocate unlimited growth in thedesignated fields, in the hopes that "super intelligent" machines mightone day colonise the universe. It is this "cosmic"view of history, in which the fate of one single species, on one singleplanet, is seen as insignificant next to the fate of the known universe, that gives the Cosmists their name.
Terrans on the other hand, will have a more "terrestrial"Earth-centred view, in which the fate of the Earth and its species(like humanity) are seen as being all-important. To Terrans, a futurewithout humans is to be avoided at all costs, as it would represent theworst-case scenario. As such, Terrans will find themselves unable toignore the possibility that super intelligent machines might one daycause the destruction of the human race -- being very immenselyintelligent and so cosmically inclined, these artilect machines mayhave no more moral or ethical difficulty in exterminating humanity thanhumans do in using medicines to cure diseases. So, Terrans will seethemselves as living during the closing of a window of opportunity, todisable future artilects before they are built, after which humans willno longer have a say in the affairs of intelligent machines.
It is these two extreme ideologies which de Garis believes mayherald a new world war, wherein one group with a 'grand plan' (theCosmists) will be rabidly opposed by another which feels itself to beunder deadly threat from that plan (the Terrans). The factions, hepredicts, may eventually war to the death because of this, as theTerrans will come to view the Cosmists as "arch-monsters" whenthey begin seriously discussing acceptable risks, and the probabilitiesof large percentages of Earth-based life going extinct. In response tothis, the Cosmists will come to view the Terrans as being reactionaryextremists, and will stop treating them and their ideas seriously,further aggravating the situation, possibly beyond reconciliation.
Throughout his book, de Garis states that he is ambivalent aboutwhich viewpoint he ultimately supports, and attempts to make convincingcases for both sides. He elaborates towards the end of the book thatthe more he thinks about it, the more he feels like a Cosmist, becausehe feels that despite the horrible possibility that humanity mightultimately be destroyed, perhaps inadvertently or at leastindifferently, by the artilects, he cannot ignore the fact that thehuman species is just another link in the evolutionary chain, and mustgo extinct in their current form anyway, whereas the artilects couldvery well be the next link in that chain and therefore would beexcellent candidates to carry the torch of science and explorationforward into the rest of the universe.
He relates a morally isomorphic scenario in which extraterrestrialintelligences visit the earth three billion years ago and discover twodomains of life living there, one domain which is older but simplerand contemporarily dominant, but which upon closer study appears to beincapable of much further evolutionary development; and one youngerdomain which is struggling to survive, but which upon further studydisplays the potential to evolve into all the varieties of lifeexisting on the Earth today, including humanity, and then queries thereader as to whether they would feel ethically compelled to destroy thedominant domain of life to ensure the survival of the younger one, orto destroy the younger one in order to ensure the survival of the olderand more populous domain which was "there first." He states that hebelieves that, like himself, most of the public would feel torn or atleast ambivalent about the outcome of artilects at first, but that asthe technology advanced, the issue would be forced and most would feelcompelled to choose a side, and that as such the public consciousnessof the coming issue should be raised now so that society can choose, hopefully before the factions becomes irreconcilably polarised, which outcome it prefers.
de Garis relates that "just out of curiosity, I asked Kevin (Warwick)whether he was a Terran or a Cosmist. He said he was against the ideaof artilects being built (i.e., he is Terran). I was surprised, andfelt a shiver go up my spine. That moment reminded me of a biography ofLenin that I had read in my 20s in which the Bolsheviks and the Mensheviks first started debating the future government of Russia. What began as an intellectual difference ended up as a Russian civil war after 1917 between the white and the red Russians."[14]
It should actually be noted, however, that Warwick is really not aTerran at all, but would in fact be a member of another party de Garispredicts will emerge between the two.[citation needed] He colloquially refers to this third party as "Cyborgians",because they will not be opposed to artilects as such, but they willdesire to personally participate in the artilect colonisation of theuniverse, rather than fall into obsolescence. They will seek to become artilects by gradually merging themselves with machines, which is the main focus of Professor Warwick's cybernetics research.
Quotes- Humans should not stand in the way of a higher form ofevolution. These machines are godlike. It is human destiny to createthem.
- I believe that the ideological disagreements between these twogroups on this issue will be so strong, that a major "artilect" war,killing billions of people, will be almost inevitable before the end ofthe 21st century.[2]:234
— speaking in 2005 of the Cosmist/Terran conflict. - Twenty years from now, the author envisages the brain builderindustry as being one of the world's top industries, comparable withoil, automobile, and construction.[1]
— prediction made in 1996. |
楼主: coredump
发表于 2-7-2009 13:06:20
如果每个研究者，都认为
楼主
