User:Mokidijohnayobami/sandbox

From Wikipedia, the free encyclopedia

Ore Oríkĕ itetisi[edit]

Part of a series on
Artificial intelligence
Major goals
Approaches
Applications
Philosophy
History
Glossary

Oye itetisi atọwọda ọrẹ (tun AI tabi FAI) jẹ oye itetisi gbogbogbo atọwọda (AGI) ti yoo ni ipa rere (aiṣedeede) lori ẹda eniyan tabi o kere ju ni ibamu pẹlu awọn ire eniyan tabi ṣe alabapin si imudara ilọsiwaju ti ẹda eniyan. O jẹ apakan ti awọn ilana iṣe ti oye atọwọda ati pe o ni ibatan pẹkipẹki si awọn ilana iṣe ẹrọ. Lakoko ti ilana iṣe ẹrọ jẹ ibakcdun pẹlu bii aṣoju ti oye atọwọda yẹ ki o huwa, iwadii itetisi atọwọda ọrẹ wa ni idojukọ lori bii o ṣe le mu ihuwasi yii ṣiṣẹ ni adaṣe ati rii daju pe o ni ihamọ to.

Etymology ati lilo[edit]

Ọrọ naa jẹ ipilẹṣẹ nipasẹ Eliezer Yudkowsky[1], ẹniti o mọ julọ fun sisọ imọran naa,[2][3] lati jiroro lori awọn aṣoju atọwọda ti o ni oye ti o ni igbẹkẹle ti o ṣe imuse awọn iye eniyan. Stuart J. Russell ati Peter Norvig's asiwaju atọwọdọwọ iwe kika, Artificial Intelligence: A Modern Approach, apejuwe awọn agutan:

Yudkowsky (2008) lọ sinu alaye diẹ sii nipa bi o ṣe le ṣe apẹrẹ AI Ọrẹ kan. O sọ pe ọrẹ (ifẹ lati ma ṣe ipalara fun eniyan) yẹ ki o ṣe apẹrẹ lati ibẹrẹ, ṣugbọn pe awọn apẹẹrẹ yẹ ki o mọ mejeeji pe awọn apẹrẹ ti ara wọn le jẹ abawọn, ati pe robot yoo kọ ẹkọ ati dagbasoke ni akoko pupọ. Nitorinaa ipenija naa jẹ ọkan ninu apẹrẹ ẹrọ-lati ṣalaye ẹrọ kan fun idagbasoke awọn eto AI labẹ eto awọn sọwedowo ati awọn iwọntunwọnsi, ati lati fun awọn iṣẹ ṣiṣe awọn ọna ṣiṣe ti yoo jẹ ọrẹ ni oju iru awọn ayipada.

John Mokidi, An I.T Specialist

'Ọrẹ' ni a lo ni aaye yii gẹgẹbi imọ-ọrọ imọ-ẹrọ, o si yan awọn aṣoju ti o wa ni ailewu ati iwulo, kii ṣe awọn ti o jẹ “ọrẹ” ni itumọ ọrọ-ọrọ. Agbekale naa ni akọkọ ti a pe ni ọrọ ti awọn ijiroro ti awọn aṣoju atọwọda ti o ni ilọsiwaju ti ara ẹni ti o nyara gbamu ni oye, lori awọn aaye pe imọ-ẹrọ arosọ yii yoo ni ipa nla, iyara, ati iṣoro-lati-ṣakoso lori awujọ eniyan.

Awọn lilo ti Ore Oríkĕ oye ni Informatics ati Public Administration[edit]

Ṣiṣe Ipinnu Iwa ni Isakoso Awujọ: Awọn ilana FAI le ṣe itọsọna idagbasoke awọn eto AI ti a lo ni iṣakoso gbogbogbo lati rii daju ṣiṣe ipinnu ihuwasi. Eyi pẹlu awọn ero bii ododo, iṣiro, ati akoyawo ni imuṣiṣẹ ti awọn algoridimu AI fun awọn iṣẹ gbogbogbo.[4]

Sihin ati Awọn ọna AI ti o ṣalaye: Awọn ipilẹ AI ọrẹ le ṣee lo lati ṣẹda awọn eto AI ni awọn alaye ti o han gbangba ati alaye. Eyi ṣe pataki ni iṣakoso gbogbogbo, nibiti awọn oluṣe ipinnu nilo lati loye ero lẹhin awọn iṣeduro tabi awọn ipinnu idari AI.[5]

Ifowosowopo Eda Eniyan-AI fun Atilẹyin Ipinnu: Ọrẹ AI le ṣe alekun ifowosowopo eniyan-AI ni iṣakoso gbogbogbo nipa fifun awọn irinṣẹ atilẹyin ipinnu. Eyi pẹlu ṣiṣe apẹrẹ awọn ọna ṣiṣe AI ti o ni ibamu pẹlu oye eniyan dipo ki o rọpo rẹ, ṣiṣe idagbasoke ajọṣepọ ti o munadoko diẹ sii ni awọn ilana ṣiṣe ipinnu.[6]

Ilọkuro Ewu ni Awọn ọna AI: Awọn ipilẹ AI ọrẹ le jẹ ohun elo ni idamo ati idinku awọn eewu ti o ni nkan ṣe pẹlu awọn eto AI ni awọn alaye. Eyi jẹ pataki ni pataki ni iṣakoso gbogbogbo, nibiti awọn aṣiṣe tabi aibikita ni ṣiṣe ipinnu AI le ni awọn abajade to ṣe pataki.[7]

Ifaramọ si Awọn Ilana Ofin ati Ilana: Awọn ilana FAI le ṣe itọsọna idagbasoke awọn eto AI lati rii daju ibamu pẹlu awọn ilana ofin ati ilana ni iṣakoso gbogbo eniyan. Eyi pẹlu awọn ero fun awọn ofin aabo data, awọn ẹtọ eniyan, ati awọn ilana miiran ti o yẹ.

Awọn ewu ti AI aifẹ[edit]

Awọn gbongbo ti ibakcdun nipa itetisi atọwọda jẹ arugbo pupọ. Kevin LaGrandeur fihan pe awọn ewu ti o ni pato si AI ni a le rii ni awọn iwe-akọọlẹ atijọ nipa awọn iranṣẹ eda eniyan atọwọda gẹgẹbi golem, tabi awọn proto-roboti ti Gerbert ti Aurillac ati Roger Bacon. Ninu awọn itan wọnyẹn, oye pupọ ati agbara ti awọn ẹda eda eniyan ni ija pẹlu ipo wọn bi ẹrú (eyiti o jẹ nipa iseda ti a rii bi iha-eniyan), ti o si fa ija ajalu.[8]

Ni awọn akoko ode oni bi ifojusọna AI alabojuto ti n sunmọ, onimọran Nick Bostrom ti sọ pe awọn eto AI ti o ni oye pẹlu awọn ibi-afẹde ti ko ni ibamu pẹlu awọn iṣe eniyan jẹ eewu lainidii ayafi ti a ba gbe awọn igbese to gaju lati rii daju aabo ẹda eniyan. Ó sọ ọ́ lọ́nà yìí:

Ni ipilẹ o yẹ ki a ro pe 'oye oye' kan yoo ni anfani lati ṣaṣeyọri awọn ibi-afẹde eyikeyi ti o ni. Nitorinaa, o ṣe pataki pupọ pe awọn ibi-afẹde ti a fi fun u, ati gbogbo eto iwuri rẹ, jẹ 'ọrẹ eniyan.'

Ni ọdun 2008 Eliezer Yudkowsky pe fun ṣiṣẹda “AI ore” lati dinku eewu ti o wa lati inu oye atọwọda ilọsiwaju. Ó ṣàlàyé pé: “ AI kò kórìíra rẹ, bẹ́ẹ̀ ni kò nífẹ̀ẹ́ rẹ, ṣùgbọ́n a fi ọ́ ṣe àwọn ọ̀mùnú tí ó lè lò fún nǹkan mìíràn.”[9]

Steve Omohundro sọ pe eto AI ti o ni ilọsiwaju ti o ni ilọsiwaju yoo, ayafi ti atako ni gbangba, ṣe afihan nọmba awọn “awọn awakọ” ipilẹ, gẹgẹbi gbigba awọn orisun, titọju ara ẹni, ati ilọsiwaju ti ara ẹni ti nlọsiwaju, nitori ẹda ara ti eyikeyi awọn ọna ṣiṣe ti ibi-afẹde eyikeyi. ati pe awọn awakọ wọnyi yoo, “laisi awọn iṣọra pataki”, fa AI lati ṣafihan ihuwasi aifẹ.[10][4]

Alexander Wissner-Gross sọ pe AI ti o ni itara lati mu iwọn ominira iṣe ọjọ iwaju wọn pọ si (tabi ipa ọna entropy) ni a le gba bi ọrẹ ti o ba jẹ pe oju-ọna igbero wọn gun ju iloro kan lọ, ati aibikita ti o ba jẹ pe ipade eto wọn kuru ju iloro yẹn lọ.[11][12]

Luke Muehlhauser, kikọ fun Ile-iṣẹ Iwadi Imọye Ọgbọn Ẹrọ, ṣe iṣeduro pe awọn oniwadi ethics ẹrọ gba ohun ti Bruce Schneier ti pe ni “ero aabo”: Dipo ki o ronu nipa bi eto kan yoo ṣe ṣiṣẹ, fojuinu bi o ṣe le kuna. Fun apẹẹrẹ, o daba paapaa AI kan ti o ṣe awọn asọtẹlẹ deede ati ibaraẹnisọrọ nipasẹ wiwo ọrọ le fa ipalara ti a ko pinnu.[13]

Ni 2014, Luke Muehlhauser ati Nick Bostrom ṣe afihan iwulo fun 'ore AI';[14] sibẹsibẹ, awọn isoro ni nse a 'ore' superintelligence, fun apẹẹrẹ nipasẹ siseto counterfactual iwa ero, jẹ akude.[15][16]

Awọn ọna miiran[edit]

Steve Omohundro ti dabaa ọna “scaffolding” kan si aabo AI, ninu eyiti iran AI ti o ni aabo ti o ni idaniloju ṣe iranlọwọ lati kọ iran ti o ni aabo ti o tẹle.[17]

Seth Baum ṣe ariyanjiyan pe idagbasoke ti ailewu, itetisi atọwọda ti o ni anfani lawujọ tabi itetisi gbogbogbo atọwọda jẹ iṣẹ ti imọ-jinlẹ awujọ ti awọn agbegbe iwadii AI, ati nitorinaa o le ni idiwọ nipasẹ awọn igbese extrinsic ati iwuri nipasẹ awọn iwọn inu. Awọn iwuri inu inu le ni okun nigbati awọn ifiranṣẹ ba tunmọ pẹlu awọn olupilẹṣẹ AI; Baum ṣe ariyanjiyan pe, ni idakeji, “awọn ifiranṣẹ ti o wa tẹlẹ nipa AI anfani ko nigbagbogbo ni ipilẹ daradara”. Baum n ṣe agbero fun “awọn ibatan ifowosowopo, ati didimu rere ti awọn oniwadi AI” ati awọn iṣọra lodi si sisọ awọn oniwadi AI bi “ko fẹ (ni) lati lepa awọn apẹrẹ anfani”.[18]

Ninu iwe rẹ Human Compatible, oluwadi AI Stuart J. Russell ṣe atokọ awọn ilana mẹta lati ṣe itọsọna idagbasoke awọn ẹrọ ti o ni anfani. Ó tẹnu mọ́ ọn pé àwọn ìlànà yìí kò túmọ̀ sí pé kí wọ́n ṣàdéhùn sára àwọn ẹ̀rọ náà; dipo, ti won ti wa ni ti a ti pinnu fun eda eniyan Difelopa. Awọn ilana jẹ bi atẹle:[19]:173

1. Ero nikan ti ẹrọ ni lati mu riri ti awọn ayanfẹ eniyan pọ si.

2. Ẹrọ naa ko ni idaniloju nipa kini awọn ayanfẹ wọnyẹn.

3. Ipilẹ orisun alaye nipa awọn ayanfẹ eniyan ni ihuwasi eniyan.

Awọn "awọn ayanfẹ" Russell n tọka si "ni gbogbo-apapọ; wọn bo ohun gbogbo ti o le bikita nipa, lainidii jina si ojo iwaju."[20]: 173  Bakanna, "ihuwasi" pẹlu eyikeyi yiyan laarin awọn aṣayan,[20]: 177  ati aidaniloju jẹ iru pe diẹ ninu awọn iṣeeṣe, eyi ti o le jẹ ohun kekere, gbọdọ wa ni sọtọ si gbogbo ọgbọn ṣee ṣe ààyò eniyan.[20]: 201

Eto imulo gbogbo eniyan[edit]

James Barrat, onkọwe ti Ipinnu Ipari Wa, daba pe "ijọṣepọ-ikọkọ ti gbogbo eniyan ni lati ṣẹda lati mu awọn oniṣẹ A.I. jọ lati pin awọn ero nipa aabo-ohun kan bi International Atomic Energy Agency, ṣugbọn ni ajọṣepọ pẹlu awọn ile-iṣẹ." O rọ awọn oniwadi AI lati pe apejọ kan ti o jọra si Apejọ Asilomar lori DNA Recombinant, eyiti o jiroro awọn ewu ti imọ-ẹrọ imọ-ẹrọ.[17]

John McGinnis ṣe iwuri fun awọn ijọba lati mu iyara iwadii AI ore. Nitoripe awọn aaye ibi-afẹde ti AI ọrẹ ko jẹ olokiki, o ni imọran awoṣe kan ti o jọra si Awọn ile-iṣẹ Ilera ti Orilẹ-ede, nibiti “Awọn panẹli atunyẹwo ẹlẹgbẹ ti kọnputa ati awọn onimọ-jinlẹ oye yoo ṣaju nipasẹ awọn iṣẹ akanṣe ati yan awọn ti a ṣe apẹrẹ mejeeji lati ṣe ilosiwaju AI ati ni idaniloju pe iru awọn ilọsiwaju bẹẹ yoo wa pẹlu awọn aabo ti o yẹ. ” McGinnis ni imọlara pe atunyẹwo ẹlẹgbẹ dara julọ “ju ilana lati koju awọn ọran imọ-ẹrọ ti ko ṣee ṣe lati mu nipasẹ awọn aṣẹ ijọba”. McGinnis ṣe akiyesi pe igbero rẹ duro ni idakeji si ti Ile-iṣẹ Iwadi Imọye Ọgbọn ẹrọ, eyiti o ni ero gbogbogbo lati yago fun ilowosi ijọba ni AI ọrẹ.[21]

Lodi[edit]

See also: Technological singularity § Criticisms

Diẹ ninu awọn alariwisi gbagbọ pe ipele AI ti eniyan ati oye ko ṣeeṣe, ati pe nitorinaa ore AI ko ṣeeṣe. Kikọ ni Olutọju naa, Alan Winfield ṣe afiwe oye itetisi atọwọda ipele eniyan pẹlu irin-ajo yiyara-ju ina ni awọn ofin iṣoro, o sọ pe lakoko ti a nilo lati “ṣọra ati mura” fun awọn ipin ti o kan, a “ko nilo lati ṣe akiyesi” nipa awọn ewu ti oye oye.[22] Boyles ati Joaquin, ni ida keji, jiyan pe Luke Muehlhauser ati imọran Nick Bostrom lati ṣẹda AIs ore dabi ẹnipe o buruju. Eyi jẹ nitori pe Muehlhauser ati Bostrom dabi pe wọn ni imọran pe awọn ẹrọ ti o ni oye ni a le ṣe eto lati ronu ni ilodi si nipa awọn iwulo iwa ti eniyan iba ti ni.[14] Ninu àpilẹkọ kan ninu AI & Society, Boyles ati Joaquin ṣetọju pe iru AI kii yoo jẹ ọrẹ ti o ṣe akiyesi awọn atẹle wọnyi: iye ailopin ti awọn ipo atako iṣaaju ti yoo ni lati ṣe eto sinu ẹrọ kan, iṣoro ti owo sisan jade ṣeto ti iwa. awọn iye-iyẹn ni, awọn ti o dara julọ ju eyi ti eniyan ni ni lọwọlọwọ, ati asopọ ti o han gbangba laarin awọn iṣaaju ti o lodi ati iye ti o dara julọ nitori abajade.[15]

Àwọn onímọ̀ ọgbọ́n orí kan sọ pé àwọn aṣojú “òye” nítòótọ́, yálà àtọwọ́dá tàbí ènìyàn, yóò jẹ́ onínúure nípa ti ara; ni wiwo yii, awọn aabo mọọmọ ti a ṣe apẹrẹ lati gbejade AI ọrẹ le jẹ ko wulo tabi paapaa ipalara.[23] Awọn alariwisi miiran beere boya o ṣee ṣe fun oye atọwọda lati jẹ ọrẹ. Adam Keiper ati Ari N. Schulman, awọn olutọsọna ti iwe iroyin imọ-ẹrọ The New Atlantis, sọ pe kii yoo ṣee ṣe lati ṣe iṣeduro ihuwasi “ọrẹ” lailai ni AI nitori awọn iṣoro ti idiju ihuwasi kii yoo fun awọn ilọsiwaju sọfitiwia tabi pọ si ni agbara iširo. Wọn kọ pe awọn ibeere lori eyiti awọn imọ-jinlẹ AI ọrẹ da lori iṣẹ “nikan nigbati ẹnikan ko ni awọn agbara nla ti asọtẹlẹ nikan nipa iṣeeṣe ti awọn abajade ti o ṣeeṣe ti ọpọlọpọ, ṣugbọn idaniloju ati ipohunpo lori bii ẹnikan ṣe ni iye awọn abajade oriṣiriṣi.[24]

See also[edit]


Awọn itọkasi[edit]

  1. ^ Tegmark, Max (2014). "Life, Our Universe and Everything". Our Mathematical Universe: My Quest for the Ultimate Nature of Reality (First ed.). Knopf Doubleday Publishing. ISBN 9780307744258. Its owner may cede control to what Eliezer Yudkowsky terms a "Friendly AI,"...
  2. ^ Russell, Stuart; Norvig, Peter (2009). Artificial Intelligence: A Modern Approach. Prentice Hall. ISBN 978-0-13-604259-4.
  3. ^ Leighton, Jonathan (2011). The Battle for Compassion: Ethics in an Apathetic Universe. Algora. ISBN 978-0-87586-870-7.
  4. ^ a b Bostrom, Nick (July 3, 2014). "Chapter 7". Superintelligence: Paths, Dangers, Strategies. Oxford University Press. p. 352pp. ISBN 978-0199678112.{{cite book}}: CS1 maint: date and year (link)
  5. ^ Doshi-Velez, Finale. n.d. “Towards A Rigorous Science of Interpretable Machine Learning.” arXiv.Org. https://arxiv.org/abs/1702.08608.
  6. ^ Brynjolfsson, Erik; McAfee, Andrew (January 25). The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies (2nd ed.). W.W. Norton & Company. ISBN 978-0393350647. {{cite book}}: Check date values in: |date= and |year= / |date= mismatch (help)
  7. ^ Amodei,, Dario; Olah, Chris; Steinhardt, Jacob; Christiano, Paul; Schulman, John; Mané, Dan (21, June 2016). "Concrete Problems in AI Safety". {{cite journal}}: Check date values in: |date= (help); Cite journal requires |journal= (help)CS1 maint: extra punctuation (link)
  8. ^ Kevin LaGrandeur (2011). "The Persistent Peril of the Artificial Slave". Science Fiction Studies. 38 (2): 232. doi:10.5621/sciefictstud.38.2.0232. ISSN 0091-7729.
  9. ^ Yudkowsky, Eliezer (2008-07-03), "Artificial Intelligence as a positive and negative factor in global risk", Global Catastrophic Risks, Oxford University Press, retrieved 2023-12-05
  10. ^ Omohundro, Stephen M. (2018-07-27), "The Basic AI Drives", Artificial Intelligence Safety and Security, First edition. | Boca Raton, FL : CRC Press/Taylor & Francis Group, 2018.: Chapman and Hall/CRC, pp. 47–55, ISBN 978-1-351-25138-9, retrieved 2023-12-05{{citation}}: CS1 maint: location (link)
  11. ^ "How Skynet Might Emerge From Simple Physics". Gizmodo. 2013-04-26. Retrieved 2023-12-05.
  12. ^ Wissner-Gross, A. D.; Freer, C. E. (2013-04-19). "Causal Entropic Forces". Physical Review Letters. 110 (16). doi:10.1103/physrevlett.110.168702. ISSN 0031-9007.
  13. ^ Muehlhauser, Luke (2013-08-01). "AI Risk and the Security Mindset". Machine Intelligence Research Institute. Retrieved 2023-12-05.
  14. ^ a b Muehlhauser, Luke; Bostrom, Nick (2013-12-17). "WHY WE NEED FRIENDLY AI". Think. 13 (36): 41–47. doi:10.1017/s1477175613000316. ISSN 1477-1756.
  15. ^ a b Boyles, Robert James M.; Joaquin, Jeremiah Joven (2019-07-23). "Why friendly AIs won't be that friendly: a friendly reply to Muehlhauser and Bostrom". AI & SOCIETY. 35 (2): 505–507. doi:10.1007/s00146-019-00903-0. ISSN 0951-5666.
  16. ^ Chan, Berman (2020-03-04). "The rise of artificial intelligence and the crisis of moral passivity". AI & SOCIETY. 35 (4): 991–993. doi:10.1007/s00146-020-00953-9. ISSN 0951-5666.
  17. ^ a b Magazine, Smithsonian; Hendry, Erica R. "What Happens When Artificial Intelligence Turns On Us?". Smithsonian Magazine. Retrieved 2023-12-05.
  18. ^ Baum, Seth D. (2016). "On the Promotion of Safe and Socially Beneficial Artificial Intelligence". SSRN Electronic Journal. doi:10.2139/ssrn.2816323. ISSN 1556-5068.
  19. ^ Russell, Stuart J. (Stuart Jonathan) (2019). Human compatible : artificial intelligence and the problem of control. Internet Archive. [New York, New York?] : Viking. ISBN 978-0-525-55861-3.
  20. ^ a b c Russell, Stuart (October 8, 2019). Human Compatible: Artificial Intelligence and the Problem of Control. United States: Viking. ISBN 978-0-525-55861-3. OCLC 1083694322.
  21. ^ McGinnis, John O. (Summer 2010). "Accelerating AI". Northwestern University Law Review. 104 (3): 1253–1270. Archived from the original on 1 December 2014. Retrieved 16 July 2014.
  22. ^ Winfield, Alan (9 August 2014). "Artificial intelligence will not turn into a Frankenstein's monster". The Guardian. Archived from the original on 17 September 2014. Retrieved 17 September 2014.
  23. ^ Kornai, András (2014-05-15). "Bounding the impact of AGI". Journal of Experimental & Theoretical Artificial Intelligence. 26 (3): 417–438. doi:10.1080/0952813x.2014.895109. ISSN 0952-813X.
  24. ^ Keiper, Adam; Schulman, Ari N. (Summer 2011). "The Problem with 'Friendly' Artificial Intelligence". The New Atlantis. No. 32. pp. 80–89. Archived from the original on 2012-01-15. Retrieved 2012-01-16.

[1]

Ita Links[edit]

  1. ^ Brynjolfsson, Erik, and Andrew McAfee. 2015. “The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies.” Choice Reviews Online 52 (06): 52–3201. https://doi.org/10.5860/choice.184834.
Retrieved from "https://en.wikipedia.org/w/index.php?title=User:Mokidijohnayobami/sandbox&oldid=1195968143"