Bugu da ari, sharhin
Wannan sashe da aka tsara za a yi amfani a matsayin tunani, maimakon a karanta a matsayin labari.
- Gabatarwa (Sashe 5.1)
Mass haɗin gwiwar blends ideas daga jama'a kimiyya, dandazon, kuma gama hankali. Citizen kimiyya yawanci yana nufin shafe "yan" (watau, wadanda ba masana kimiyya) a cikin kimiyya tsari (Crain, Cooper, and Dickinson 2014) . Taron yawanci yana nufin shan matsala yawanci warware cikin wani shiri kuma a maimakon samuwan kaya daga waje da shi a wani taron (Howe 2009) . Collective m yawanci yana nufin kungiyoyin mutane aiki tare a hanyoyi da ze fasaha (Malone and Bernstein 2015) . Nielsen (2012) shi ne mai ban mamaki littafin-tsawon gabatarwar a cikin ikon taro haɗin gwiwar kimiyya da bincike.
Akwai su da yawa iri taro haɗin gwiwar da ba su dace neatly a cikin sassa uku da na samarwa, kuma ina ganin uku cancanci musamman da hankali, domin su zama da amfani a social bincike a wasu batu. Daya misali ne Hasashen kasuwanni, inda mahalarta saya da cinikayya kwangila da aka cancanci samun ceto bisa sakamakon da faruwa a cikin duniya (Wolfers and Zitzewitz 2004; Arrow et al. 2008) . Tsinkaya kasuwanni sukan yi amfani da kamfanonin da gwamnatoci ga kiyasin da tsinkaya kasuwanni sun kuma an yi amfani da zamantakewa masu bincike zuwa hango ko hasashen replicability na buga karatu a tunani (Dreber et al. 2015) .
A karo na biyu misali da cewa ba ya shige da kyau a cikin ta categorization makirci ne PolyMath aikin, inda masu bincike hada kai ta yin amfani da blogs da wikis tabbatar da sabon ilimin lissafi theorems (Gowers and Nielsen 2009; Cranshaw and Kittur 2011; Nielsen 2012; Kloumann et al. 2016) . The PolyMath shiri ne a wasu hanyoyi kama da Netflix Prize, amma a PolyMath aikin mahalarta more rayayye gina a kan m mafita wasu.
A uku misali da cewa ba ya shige da kyau a cikin ta categorization makirci ne lokaci-dogara mobilizations kamar tsaron Advanced Research Projects Agency (DARPA) Network Challenge (ie, da Red balan-balan Challenge). Don ƙarin a kan wadannan lokaci m mobilizations ganin Pickard et al. (2011) , Tang et al. (2011) , da kuma Rutherford et al. (2013) .
- Human ƙidãyar (Sashe 5.2)
Kalmar "mutum ƙidãyar" ya zo daga aikin da kwamfuta masana kimiyya, da kuma fahimtar mahallin bayan wannan bincike zai inganta ikon tara daga matsalolin da zai yi amenable zuwa gare shi. Domin wasu ayyuka, kwakwalwa ne wuce yarda m da damar zuwa yanzu wucewa har gwani mutane. Alal misali, a dara, kwakwalwa iya doke har ma da mafi kyau grand Masters. Amma, wannan yana kasa da yaba da zamantakewa masana kimiyya-ga sauran ayyuka, kwakwalwa ne a zahiri ya fi mutane. A wasu kalmomin, a yanzu ku ne mafi alhẽri daga ko da ya fi sophisticated kwamfuta a wasu ayyuka shafe aiki na images, video, audio, da kuma rubutu. Kamar wancan-kamar yadda aka kwatanta da mai ban mamaki XKCD zane mai ban dariya-akwai ayyuka da cewa su ne m for kwakwalwa da kuma m ga mutane, amma akwai kuma ayyuka da cewa su ne m for kwakwalwa da kuma sauki ga mutane (Figure 5.13). Computer masana kimiyya aiki a kan wadannan wuya-ga-kwakwalwa-sauki-ga-adam ayyuka, sabili da haka, ya gane cewa su iya sun hada da mutane a cikin mai aiki da na'urar kwamfuta tsari. Ga yadda Luis von Ahn (2005) ya bayyana mutum ƙidãyar a lõkacin da ya fara buga da kalmar a cikin dissertation: "a yayi domin amfani mutum aiki ikon warware matsalolin da kwakwalwa ba zai iya yet shirya."
Adadi 5.13: Ga wasu ayyuka kwakwalwa ne mai ban mamaki, wucewa da ikon da mutum masana. Amma, ga sauran ayyuka, talakawa mutane na iya outperform har sophisticated kwakwalwa tsarin. Large sikelin matsalolin da unsa ayyuka da cewa su ne m for kwakwalwa da kuma sauki ga mutane ne da-dace domin mutum ƙidãyar. Used bisa ga sharuddan da aka bayyana a nan: http://xkcd.com/license.html
By wannan definition FoldIt-da na bayyana a cikin sashen bude kira-a iya dauke da wani mutum ƙidãyar aikin. Duk da haka, na zabi don rarrabesu FoldIt matsayin bude kira domin shi bukatar na musamman basira kuma yana daukan mafi kyau bayani da gudummawar maimakon ta amfani da tsaga-tambaya-hada dabarun.
Domin mai kyau littafin tsawon jiyya na mutum ƙidãyar, a cikin mafi general ji da kalmar, gani Law and Ahn (2011) . Babi na 3 na Law and Ahn (2011) yana da ban sha'awa game da more hadaddun hada matakai fiye da wadanda a cikin wannan babi.
Kalmar "tsaga-tambaya-hada" da aka yi amfani da Wickham (2011) bayyana a dabarun domin ilimin kididdiga kwamfuta, amma daidai kama da aiwatar da yawa mutum ƙidãyar ayyukan. The tsaga-tambaya-hada dabarun ne kama da MapReduce tsarin raya a Google (Dean and Ghemawat 2004; Dean and Ghemawat 2008) .
Biyu wayo mutum ƙidãyar ayyukan da na ba su da tsawo domin tattauna ne Esp Game (Ahn and Dabbish 2004) da kuma reCAPTCHA (Ahn et al. 2008) . Duka wadannan ayyukan samu m hanyoyin da za a tilasta mahalarta su samar da tasirin a kan images. Duk da haka, biyu daga cikin wadannan ayyukan ma tashe da'a tambayoyi domin, sabanin Galaxy Zoo, mahalarta a cikin Esp Game da reCAPTCHA bai san yadda za su data ake amfani (Lung 2012; Zittrain 2008) .
Wahayi zuwa gare ta Esp Game, da yawa masu bincike suka yi ƙoƙarin samar da wasu "games da wani dalili" (Ahn and Dabbish 2008) (watau "mutum na tushen ƙidãyar games" (Pe-Than, Goh, and Lee 2015) ) cewa zai iya zama amfani da su warware da dama sauran matsaloli. Abin da wadannan "games da wani dalili" da a kowa ne cewa sun yi kokarin yin ayyuka da hannu a mutum ƙidãyar m. Saboda haka, yayin da Esp Game hannun jari guda tsaga-tambaya-hada tsarin da Galaxy Zoo, shi ya bambanta da irin yadda mahalarta suna m-fun vs. so ya taimake kimiyya.
My bayanin Galaxy Zoo fa, tã a kan Nielsen (2012) , Adams (2012) , Clery (2011) , da kuma Hand (2010) , da kuma na gabatar na gudanar da bincike a raga na Galaxy Zoo aka Sauki. Don ƙarin a kan tarihin galaxy rarrabuwa a ilmin taurari da kuma yadda Galaxy Zoo ci gaba da wannan al'ada, ga Masters (2012) da kuma Marshall, Lintott, and Fletcher (2015) . Building a Galaxy Zoo, da masu bincike kammala Galaxy Zoo 2 wanda tattara fiye da miliyan 60 more hadaddun morphological sukayi fassara daga masu sa kai (Masters et al. 2011) . Bugu da ari, sun branched daga cikin matsalolin waje na galaxy ilimin halittar jiki ciki har da binciko surface da watã, neman taurari, da kuma sauya rubũtun tamkar tsohon takardun. Currently, dukan ayyukan da ake tattara a www.zooniverse.org (Cox et al. 2015) . Daya daga cikin ayyukan-Snapshot Serengeti-samar da shaida cewa Galaxy Zoo-type image rarrabuwa ayyukan kuma za a iya yi wa muhalli bincike (Swanson et al. 2016) .
Domin bincike da shirin yin amfani da wani micro-aiki aiki kasuwa (misali, Amazon Mechanical Turk) ga wani mutum ƙidãyar aikin, Chandler, Paolacci, and Mueller (2013) da kuma Wang, Ipeirotis, and Provost (2015) bayar da shawara mai kyau a kan aiki zane da kuma sauran related al'amurran da suka shafi.
Masu bincike sha'awar samar da abin da Na kira biyu ƙarni mutum ƙidãyar tsarin (misali, tsarin da suke amfani da mutum tasirin horar da wata na'ura ilmantarwa model) zai zama sha'awar Shamir et al. (2014) (ga wani misali ta amfani da audio) da kuma Cheng and Bernstein (2015) . Har ila yau, wadannan ayyukan za a iya yi tare da bude kira, inda masu bincike gasa don ƙirƙirar na'ura ilmantarwa model da mafi girma gaibu yi. Alal misali, Galaxy Zoo tawagar gudu an bude kira kuma sami wani sabon tsarin kula da outperformed daya ci gaba a Banerji et al. (2010) . ganin Dieleman, Willett, and Dambre (2015) domin cikakkun bayanai.
- Open kira (Sashe 5.3)
Open kira ne ba sabon. A gaskiya ma, daya daga cikin mafi yawan sanannun bude kira Dates a mayar da 1714 a lokacin da kasar Birtaniya ta majalisar halitta The Longitude Prize for kowa da zai iya samar da hanyar sanin da longitude da wani jirgin a teku. Matsalar maka batir da dama daga cikin mafi girma da masana kimiyya na days, ciki har da Ishaku Newton, da kuma lashe bayani da aka ƙarshe ƙaddamar da wani clockmaker daga karkara suka matso kusa da matsala daban daga masana kimiyya suka mayar da hankali a kan wani bayani da zai ko ta yaya unsa ilmin taurari (Sobel 1996) . Kamar yadda wannan misali ya nuna, daya dalilin cewa bude kira ana zaton aiki sosai ne cewa su samar da damar yin amfani da mutane daban-daban tare ra'ayoyi da kwarewa (Boudreau and Lakhani 2013) . Dubi Hong and Page (2004) da kuma Page (2008) domin more on darajar bambancin a matsalar warwarewa.
Kowace daga cikin bude kira lokuta a cikin babi na bukatar a bit of kara bayani ga dalilin da ya sa shi ne da mulkin a cikin wannan category. Na farko, hanya daya da na rarrabe tsakanin mutum ƙidãyar da kuma bude kira ayyukan ne ko fitarwa ne talakawan dukan mafita (mutum ƙidãyar) ko mafi kyau bayani (bude kira). The Netflix Prize da ɗan tricky a wannan batun, domin mafi kyau bayani ya juya a kira su da wani nagartaccen talakawan na mutum mafita, an kusata kira wani gungu bayani (Bell, Koren, and Volinsky 2010; Feuerverger, He, and Khatri 2012) . Daga hangen zaman gaba da Netflix, duk da haka, duk suka yi ta yi shi sama da mafi kyaun bayani.
Na biyu, da wasu fassarorin mutum ƙidãyar (misali, Von Ahn (2005) ), FoldIt kamata a yi la'akari da mutum ƙidãyar aikin. Duk da haka, na zabi don rarrabesu FoldIt matsayin bude kira domin shi bukatar na musamman basira kuma yana daukan mafi kyau bayani da gudummawar, maimakon ta amfani da tsaga-tambaya-hada dabarun.
A karshe, wanda zai iya bayar da hujjar cewa Peer-to-Patent misalin rarraba data collection. Na zabi a hada da shi a matsayin wani kira a bayyane saboda shi yana da hamayya-kamar tsarin da kawai mafi kyau gudunmawar da ake amfani (alhãli kuwa tare da rarraba bayanai da tarin, da ra'ayin mai kyau da kuma mummuna gudunmawar ne kasa bayyana).
Don ƙarin a kan Netflix Prize, gani Bennett and Lanning (2007) , Thompson (2008) , Bell, Koren, and Volinsky (2010) , da kuma Feuerverger, He, and Khatri (2012) . Don ƙarin on FoldIt gani, Cooper et al. (2010) , Andersen et al. (2012) , da kuma Khatib et al. (2011) . ta bayanin FoldIt fa, tã a kan kwatancin a Nielsen (2012) , Bohannon (2009) , da kuma Hand (2010) . Don ƙarin on Peer-to-Patent, gani Noveck (2006) , Bestor and Hamp (2010) , Ledford (2007) , da kuma Noveck (2009) .
Similar to sakamakon Glaeser et al. (2016) , Mayer-Schönberger and Cukier (2013) , Babi na 10 rahotanni manyan samu a cikin yawan aiki na mahalli sufetocin a New York City lokacin da dubawa an shiryar da gaibu model. A New York City, wadannan gaibu model aka gina da birnin ma'aikata, amma a wasu lokuta, wanda zai iya tunanin cewa za su iya a halitta ko inganta tare da bude kira (misali, Glaeser et al. (2016) ). Duk da haka, wanda babbar damuwa da gaibu model ake amfani da su ware albarkatun ne cewa model da m to ƙarfafa data kasance biases. Mutane da yawa masu bincike riga ya san "datti a, datti daga", kuma da gaibu model zai iya zama "nuna bambanci a, nuna bambanci ba." Dubi Barocas and Selbst (2016) da kuma O'Neil (2016) don ƙarin kan hatsarori da gaibu model gina tare da son zuciya horo data.
Daya matsalar da zai hana gwamnatoci daga yin amfani da bude gasa shi ne cewa shi bukatar data saki, wanda zai iya haifar da take hakki bayanin tsare. Domin ƙarin bayani game da tsare sirri da kuma data saka a bude kira ga Narayanan, Huey, and Felten (2016) da kuma tattaunawa a Babi na 6.
- Rarraba data tarin (Sashe 5.4)
My bayanin eBird fa, tã a kan kwatancin a Bhattacharjee (2005) da kuma Robbins (2013) . Don ƙarin a kan yadda masu bincike amfani ilimin kididdiga model to bincika eBird data gani Hurlbert and Liang (2012) da kuma Fink et al. (2010) . Don ƙarin a kan tarihin dan kimiyya a ornothology, gani Greenwood (2007) .
Don ƙarin a kan Malawi mujallolin Project, gani Watkins and Swidler (2009) da kuma Kaler, Watkins, and Angotti (2015) . Kuma more on a related aikin a Afrika ta Kudu, ganin Angotti and Sennott (2015) . Don ƙarin misalai na bincike ta amfani da data daga Malawi mujallolin Project ga Kaler (2004) da kuma Angotti et al. (2014) .
- Zayyana your own (Sashe 5.5)
My tsarin kula da miƙa zane shawara ya inductive, bisa ga misalai na ci nasara da kasa taro haɗin gwiwar ayyukan da Na ji game da. Akwai kuma wani rafi bincike ƙoƙarin tambaya more general zamantakewa m theories to zayyana online al'ummomin da ke dace da zane na taro haɗin gwiwar ayyukan, duba, ga misali, Kraut et al. (2012) .
Game da himmatuwa mahalarta, shi ne ainihin quite tricky gane daidai da ya sa mutane shiga cikin taro haɗin gwiwar ayyukan (Nov, Arazy, and Anderson 2011; Cooper et al. 2010, Raddick et al. (2013) ; Tuite et al. 2011; Preist, Massung, and Coyle 2014) . Idan ka shirya don tilasta mahalarta da biyan kan wani micro-aiki aiki kasuwa (misali, Amazon Mechanical Turk) Kittur et al. (2013) yayi wasu shawara.
Game da kunna mamaki, don ƙarin misalai na m binciken fitowa daga Zoouniverse ayyukan, duba Marshall, Lintott, and Fletcher (2015) .
Game da kasancewa da'a, wasu kyau general gabatarwa ga al'amurran da suka shafi da hannu ne Gilbert (2015) , Salehi et al. (2015) , Schmidt (2013) , Williamson (2016) , Resnik, Elliott, and Miller (2015) , da kuma Zittrain (2008) . Domin al'amurran da suka shafi musamman alaka shari'a al'amurran da suka shafi da taron ma'aikata, gani Felstiner (2011) . O'Connor (2013) bayani da tambayoyi game da da'a lura da bincike a lõkacin da matsayin da masu bincike da kuma mahalarta blur. Domin al'amurran da suka shafi alaka sharing data yayin kare participats a ɗan ƙasa kimiyya ayyukan, duba Bowser et al. (2014) . Dukansu Purdam (2014) da kuma Windt and Humphreys (2016) da wasu shawarwari game da al'amurran da suka shafi da'a a rarraba data collection. A karshe, mafi yawan ayyukan amince da gudunmawar amma ba ka marubucin bashi to mahalarta. A Foldit, da Foldit 'yan wasan sukan jera a matsayin wani marubucin (Cooper et al. 2010; Khatib et al. 2011) . A wasu bude kira ayyukan, da lashe gudummawa iya sau da yawa rubuta wata takarda da ya bayyana su da mafita (misali, Bell, Koren, and Volinsky (2010) da kuma Dieleman, Willett, and Dambre (2015) ). A cikin Galaxy Zoo iyalin ayyukan, musamman aiki da muhimmanci bayar da gudunmawa wani lokaci ana kiran su zuwa zama co-marubuta a takardunku. Alal misali, Ivan Terentev kuma Tim Matorny, biyu Radio Galaxy Zoo mahalarta daga Rasha, sun co-marubuta a kan daya daga cikin takardun da ya tashi daga wannan shiri (Banfield et al. 2016; Galaxy Zoo 2016) .
