Nhungamiro yehunyanzvi

Modero Makadhi uye Datasheets for Datasets

Kadhi remuenzaniso ipfupi, gwaro rakarongeka rinotsanangura kuti muenzaniso weAI ndewei, wakaongororwa sei, unoshanda sei kumapoka akasiyana uye mamiriro, uye kwausingafanirwe kushandiswa.

Teerera kugwaro iriAI-yakarondedzerwa · ~3 min

Pa peji ino3 min verenga
  1. Pfupiso
  2. Kudzika Kwakadzika
  3. Kuitwa muNyika yeChokwadi
  4. Ramangwana reModel Makadhi uye Datasheets yeDatasets
  5. Kuzvishandisa
  6. Ramba Uchiongorora
  7. Mibvunzo inowanzo bvunzwa

Pfupiso

Tsamba yedatasets inoita zvimwechete kune data rekudzidzisa: nei data rakaunganidzwa, rinei, rakaunganidzwa uye rakanyorwa sei, uye kuti rinofanira kuchengetedzwa sei. Aya magwaro ane basa nekuti anoregedza vatengi, vashandisi, vaongorori uye vanodzora kana hurongwa hwakakodzera chinangwa, uye mitemo yakaita seEU AI Act ikozvino inoda zvinyorwa zvakafanana.

Kudzika Kwakadzika

Makadhi emuenzaniso akakurudzirwa naMargaret Mitchell nevamwe vaaishanda navo mupepa re "Model Cards for Model Reporting", yakaratidzwa pamusangano weACM FAT * muna 2019. Zvikamu zvakakurudzirwa ndezviyero zvemuenzaniso, chinangwa chekushandiswa, zvinhu (zvakadai sehuwandu hwevanhu kana mamiriro ezvakatipoteredza), metrics, evaluation data, data yekudzidzira, kuongororwa kwehuwandu, kufunga kwetsika uye mazano, uye mazano. Pfungwa yakakosha ndeyekuongorora kwakapatsanurwa: kushuma kuita kwakaparadzana kune akakodzera mapoka kwete seavhareji yezvibodzwa zvinogona kuvanza kukundikana. Datasheets yeDatasets, naTimnit Gebru uye vaaishanda navo, yakatanga kutenderera muna 2018 uye yakazoburitswa muKukurukurirana kweACM. Inokwereta kubva kumagetsi madhata. Vagadziri veDataset vanopindura mibvunzo nezve kukurudzira, kuumbwa, kuunganidza maitiro, preprocessing uye kunyora, kushandiswa, kugovera uye kugadzirisa. Mafomati ane hukama anosanganisira zvirevo zvedata zvemitauro yedata uye makadhi edata. Makadhi eSystem, anoshandiswa neMeta uye OpenAI kubvira 2022 uye zvakanyanya kuGPT-4 muna 2023, anonyora hurongwa hwese hwakaiswa, kusanganisira kuyedza kuchengetedza uye kudzikisira, kwete modhi chete. Regulation yakashandura yakawanda yeiyi tsika kuita chinodikanwa. Iyo EU AI Act inoda zvinyorwa zvehunyanzvi zveakanyanya njodzi AI masisitimu, akaiswa muAnnex IV. Vanopa ma-general-chinangwa AI modhi vanofanira kuchengeta zvinyorwa zvehunyanzvi, kupa ruzivo kune vanopa pasi, uye kushambadza pfupiso yezvidzidzo zvemukati vachishandisa Komisheni template. Imwe pfungwa isiriyo ndeyekuti kadhi remuenzaniso rinoratidza kuti modhi yakachengeteka. Iko kuburitswa, kwete chitupa, uye kadhi rinogona kunge risina kukwana, rekare kana kunyorwa sekushambadzira. Kukosha kwayo kunotsamira paganhuriro dzechokwadi uye ongororo iyo mumwe munhu anogona kuberekazve.

Kuitwa muNyika yeChokwadi

Chipatara chinoongorora ganda-ganda rinoverenga kadhi remuenzaniso uye rinowana huchokwadi hwakataurwa zvakasiyana neganda. Inobva yafunga kumhanyisa yayo yekusimbisa kwenzvimbo isati yashandiswa.

Iyo Hugging Face modhi repository inoratidza kadhi remhando ine YAML metadata inoronga rezinesi rayo, dhatabheti rekudzidzisa, mitauro uye mabhenji mabhenji, uye vanhu vanogona kusefa mibairo yekutsvaga neminda iyoyo.

Chikwata chekutsvagisa chinoburitsa dataset yekutaura ine dhatabheti inorekodha mvumo yevatauri, mamiriro ekurekodha uye mataurirwo asina kumiririrwa.

Iyo AI lab inoburitsa kadhi system yechatbot inotsanangura tsvuku-team mhedzisiro, kuchengetedza kudzikisira uye njodzi dzasara, kupfuura iyo yekutanga modhi kadhi inovhara.

Ramangwana reModel Makadhi uye Datasheets yeDatasets

Zvinyorwa zviri kufamba kubva mukuzvidira zvakanakisa kuita kuenda kune chinodiwa chekuteerera. Iyo EU AI Act, mitemo yekutenga uye zviyero zvakaita se ISO/IEC 42001 zvese zvinotarisira marekodhi akarongeka ekuti sisitimu yakavakwa nekuedzwa sei. Mafomati anoverengeka nemuchina uye otomatiki chizvarwa angangove akajairika, kuitira kuti makadhi atariswe nekuenzaniswa. Matambudziko arikuenderera mberi ndeekuchengeta magwaro echokwadi sezvo mamodheru achigadziridzwa, kuburitsa zvakakwana nezve data rekudzidzisa pasina kufumura zvakavanzika zvekutengesa kana ruzivo rwemunhu, uye kuve nechokwadi chekuti vaverengi vanoshandisa chaizvo zvinyorwa pakusarudza.

Kuzvishandisa

Nei zvichikosha

  • Mutengo uye bhajeti. Zvisarudzo zvezvivakwa zvinotyaira kuita uye mutengo wekushandisa kwemakore.
  • Sarudzo dzakajeka. Dzidzo yehunyanzvi inobatsira zvikwata kusarudza murwi wakakodzera, kwete iwo mutsva chete.
  • Kudzora kwemhando yepamusoro. Sarudzo dzeinjiniya dziri nani dzinoderedza zviitiko zvekuvimbika mukugadzira.

Chenjera

  • Kugadzirisa imwe bhenji kunogona kuvanza yakafara system kushaya simba.
  • Infrastructure uye mari yekugadzirisa inowanzotarisirwa pasi.
  • Chengetedzo uye kucherechedzwa mapundu anogona kukura sezvo masisitimu anowedzera kuoma.

Matanho anotevera

  1. Tsanangura latency, mhando, uye mutengo zvinangwa usati waitwa.
  2. Benchmark pasi pechokwadi mutoro uye data mamiriro.
  3. Chishandiso chekutarisa zvikanganiso, kudonha, uye mushandisi maitiro.
  4. Gadzirira nzira dzekudzosera kumashure uye dzezviitiko usati wawedzera.

Ramba Uchiongorora

Tsamba yemahara

Tora iyo yemazuva ese AI pfupiso

Nhatu dzakasimbiswa dzeAI nyaya mangwanani ega ega pakati pevhiki, dzakanyorwa neChirungu chakajeka. Yemahara zvachose, hapana ads.

Imwe email zuva rega rega revhiki. Bvisa kunyoresa kamwe chete. Hatimbotengese kana kugovera kero yako.

Zviedze iwe pachako

Take the Model Cards and Datasheets for Datasets quiz

Mhinduro yepakarepo pamhinduro yega yega, uye chitupa chinogovaniswa chine ID inobatika kana wapedza kosi.

Tanga mibvunzo

Tsigira dzidzo yemahara yeAI. AI Understanding ndeye 501(c)(3) isingabatsiri - hapana ads, hapana paywall, zvachose. Ita chipo

Mibvunzo inowanzo bvunzwa

Chii chinonzi Model Cards uye Datasheets for Datasets?

Kadhi remuenzaniso ipfupi, gwaro rakarongeka rinotsanangura kuti muenzaniso weAI ndewei, wakaongororwa sei, unoshanda sei kumapoka akasiyana uye mamiriro, uye kwausingafanirwe kushandiswa. Tsamba yedatasets inoita zvimwechete kune data rekudzidzisa: nei data rakaunganidzwa, rinei, rakaunganidzwa uye rakanyorwa sei, uye kuti rinofanira kuchengetedzwa sei. Aya magwaro ane basa nekuti anoregedza vatengi, vashandisi, vaongorori uye vanodzora kana hurongwa hwakakodzera chinangwa, uye mitemo yakaita seEU AI Act ikozvino inoda zvinyorwa zvakafanana.

Chii chinangwa chikuru chekadhi remuenzaniso?

Kadhi remuenzaniso rinopa vaverengi zvavanoda kuti vaone kana modhi inokodzera kumwe kushandiswa, kusanganisira kwainoita zvisina kunaka.

Ndiani akatungamira bepa raikurudzira makadhi emuenzaniso?

Margaret Mitchell nevamwe vaaishanda navo vakaratidza "Model Cards for Model Reporting" paACM FAT* muna 2019.

Chii chinonzi disaggregated evaluation?

Kupatsanura mibairo neboka diki kunofumura kukundikana uko avhareji yese yaizovanza.

Datasheets yeDatasets inokwereta zano rayo kubva kumunda upi?

Zvinhu zvemagetsi zvinouya nemadhetabheti anotsanangura maitiro avo. Gebru nevamwe vaaishanda navo vakashandisa iyo pfungwa kune dataset.

Kadhi rehurongwa rakasiyana sei nekadhi remuenzaniso kadhi?

Masisitimu makadhi, akadai seOpenAI's eGPT-4, anotsanangura njodzi dzechigadzirwa chakaiswa, kubatana-kutsvuka uye kuchengetedza.