2.3.8 Algorithmically kunya

Abun hali a cikin manyan bayanai ba tsarin halitta bane; an kaddamar da shi ta hanyar aikin injiniya na tsarin.

Kodayake yawancin manyan kafofin watsa labarun basu da karfi saboda mutane ba su san cewa an rubuta bayanan su (sashi na 2.3.3), masu bincike kada suyi la'akari da hali a cikin wadannan sassan yanar-gizon su zama "al'amuran yanayi". daɗaɗɗen haɓaka don ƙaddamar da halayen halayen kamar danna kan tallan ko aika bayanai. Hanyoyin da manufofin masu zane-zane suke tsarawa a cikin bayanai an kira algorithmic confounding . Algorithmic rikitawa ba shi da sananne ga masana kimiyyar zamantakewa, amma wannan babbar damuwa ne tsakanin masana kimiyya. Kuma, sabanin wasu matsalolin da ke tattare da nau'ikan dijital, algorithmic confounding ne mafi yawan ganuwa.

Misali mai sauƙi na rikice-rikicen algorithmic shine gaskiyar cewa a kan Facebook akwai adadin masu amfani da kusan 20, kamar yadda Johan Ugander da abokan aiki suka gano (2011) . Masana kimiyya da ke nazarin wannan bayanan ba tare da fahimtar irin yadda Facebook ke aiki zai iya haifar da labaru da yawa game da yadda 20 ke kasancewa irin nau'in zamantakewa na sihiri ba. Abin farin cikin, Ugander da abokan aikinsa suna da cikakken fahimtar tsarin da suka samar da bayanai, kuma sun san cewa Facebook ta karfafa mutane tare da 'yan haɗin kan Facebook don yin abokantaka har sai sun kai abokai 20. Kodayake Ugander da abokan aiki ba su fa] a wannan ba a cikin takarda, wannan mahimmanci ne Facebook ya tsara don taimaka wa sababbin masu amfani su zama masu aiki. Ba tare da sanin game da wanzuwar wannan manufar ba, to, yana da sauƙi in samo kuskuren kuskure daga bayanan. A wasu kalmomi, abin mamaki da yawan mutane da ke da abokai 20 sun gaya mana game da Facebook fiye da yadda mutum yake ciki.

A cikin wannan misali na baya, rikicewar algorithmic ya haifar da wani sakamakon da ya faru wanda mai bincike mai bincike zai iya gano kuma yayi bincike a gaba. Duk da haka, akwai wani maɗaukakiyar fashewar algorithmic da ke faruwa a lokacin da masu zane na tsarin layi suna sane da abubuwan da suka shafi zamantakewa sannan kuma suyi wanka wadannan ka'idodin cikin aiki na tsarin su. Masana kimiyyar zamantakewar al'umma suna kiran wannan aiki : lokacin da ka'idar ta sauya duniya ta hanyar da zata kawo duniya cikin layi tare da ka'idar. A yanayin sauyewar algorithmic masu rikitarwa, yanayin da ke kunya da bayanai yana da wuyar ganewa.

Ɗaya daga cikin misalan samfurin da aka samar ta hanyar aiki shine haɗuwa a cikin hanyoyin sadarwar kan layi. A shekarun 1970 zuwa 1980, masu bincike sun gano cewa idan kun kasance abokai da Alice da Bob, to, Alice da Bob sun kasance abokai da juna fiye da idan sun kasance mutane biyu ba zaɓaɓɓu ba. An samo irin wannan nau'in a cikin labaran zamantakewa akan Facebook (Ugander et al. 2011) . Ta haka ne, wanda zai iya ɗauka cewa alamu na abota a kan Facebook yana nuna alamu na abokantaka na intanet, a kalla a cikin yanayin haɗuwa. Duk da haka, girman girman wucewa a cikin zane-zane na Facebook yana takaitawa ta hanyar algorithmic confounding. Wato, masana kimiyya a Facebook sun san kwarewar bincike da bincike game da yadda ake iya wucewa sannan sannan suka yi amfani da shi yadda Facebook ke aiki. Facebook yana da "Mutum Za Ka San" alama da ke nuna sababbin abokai, da kuma hanyar da Facebook ta yanke shawarar wanda ya ba da shawarar zuwa gare ku shi ne haɗuwa. Wato, Facebook zai iya bayar da shawarar cewa ka zama abokai tare da abokan abokanka. Wannan fasali yana da tasiri na karuwa da wucewa a cikin shafukan yanar gizo na Facebook; a wasu kalmomi, ka'idar transitivity ta kawo duniya cikin layi tare da ka'idodin ka'idar (Zignani et al. 2014; Healy 2015) . Saboda haka, lokacin da manyan bayanan bayanan sun fara haifar da farfadowar ka'idar zamantakewa, dole ne mu tabbata cewa ka'idar kanta ba ta gasa ba yadda tsarin yake aiki.

Maimakon yin tunani game da manyan bayanan bayanai kamar yadda suke lura da mutane a cikin yanayi na halitta, wani misali mafi kyau shine kallon mutane a cikin gidan caca. Casinos sune yanayin da aka haɓaka da kyau don tsara wasu halaye, kuma mai bincike ba zai taba yin tunanin kasancewa a cikin gidan caca don samar da taga mai zurfi cikin halin mutum. Hakika, zaku iya koyon wani abu game da halin mutum ta hanyar nazarin mutane a cikin casinos, amma idan kun yi watsi da gaskiyar cewa an kirkiro bayanan a cikin gidan caca, zaku iya samo wasu mummunan ra'ayi.

Abin takaici, yin la'akari da rikice-rikicen algorithmic yana da wuyar gaske saboda yawancin siffofi na tsarin yanar gizon sunadaran abu ne, an rubuta su da kyau, kuma suna canzawa akai-akai. Alal misali, kamar yadda zan bayyana a baya a cikin wannan babi, rikice-rikice na algorithmic wani bayani ne mai yiwuwa don rashin daidaituwa na Google Flu Trends (sashe 2.4.2), amma wannan ƙidayar ya wuya a tantance saboda ayyukan ciki na Google's search algorithm yan jarirai. Halin yanayin da ke tattare da algorithmic rikitarwa shine nau'i nau'i na drift tsarin. Algorithmic confounding yana nufin cewa ya kamata mu yi hankali game da duk wani da'awar game da halin mutum wanda ya zo ne daga tsarin dijital guda ɗaya, ko ta yaya girman.