Chii chaitika
Batsira Net Security inoshuma kuti AWS yakaburitsa pachena Deception , iyo inoedza kana ma AI mamodheru achikwanisa kusiyanisa kusakwana kwesoftware kubva kune yakachengeteka kodhi ine inorasisa maitiro ekusagadzikana. Mucherechedzo unosanganisira 14,822 samples dzinotora 16 programming mitauro uye anopfuura makumi manomwe CWE zvikamu; 9,695 inopihwa uye 5,127 yakasanganiswa mukati semasampuli asina kunyorwa.
Batsira Net Security inoshuma kuti AWS yakaongorora gumi nembiri-zvinangwa zvemhando kubva kune vashanu vanopa vachishandisa yakagadzirirwa kutenderedza manyepo. Mienzaniso yakachengeteka yebhenji ine maitiro chaiwo ekusagadzikana pamwe chete nedziviriro inodzivirira kushandiswa. Mamwe matambudziko anosiyana mamiriro ekutumirwa, seKubernetes network marongero, saka modhi inofanirwa kufunga nezve kodhi uye nharaunda yayo.
Sekureva kwemushumo, AWS yakagadzira uye yakakwenenzverwa mienzaniso inopesana nemhando dzemuganho, kusasanganisa masampuli aive ari nyore kurongedza. AWS inoti maitiro aya akadya makumi emabhiriyoni ematokeni. Iyo kambani inoburitsa pachena masampuli asi inonyima zvinyorwa zvawo; vashandisi vanoendesa kufanotaura kuAWS kuti vawane zvibodzwa zvakasimbiswa. Kwakabva hakutauri kuti kuwana zvibodzwa kunobhadhara here kana zvimwe zvinodiwa pakukodzera.
Batsira Net Chengetedzo inoshuma kuti vaongorori vakazvimiririra vanotarisa mavara vasina kuona sarudzo dzeumwe neumwe kana kufunga kwepakutanga. Sampuli dzinopokana dzinowana kumwe kuongororwa, uye nyaya dzisina kugadziriswa dzinoendeswa kune isina kuvharwa seti. AWS inoti isingasviki zvikamu zvitatu kubva muzana zvemasamples ezvibodzwa zvinoramba zvichikwikwidzwa mushure mekuongorora, paine tarisiro yevasingasviki 1% vasara ongororo yevanhu, uye inoshuma kuti hapana zvikanganiso zvekunyora mukuongorora kwemazana zana akasarudzwa zvisina tsarukano. Zvikumbiro izvi hazvina kusimbiswa zvakazvimirira pano.
Kwakabva mashoko: helpnetsecurity.com ↗
Nei zvichikosha
Mhedzisiro yacho inoratidza kuti kuita kwakasimba pakutsvaga kodhi inonyumwira haingoshandure kuita yakavimbika yenjodzi triage. Mitengo yepamusoro yenhema-yakanaka inogona kuremedza zvikwata zvekuchengetedza uye kuderedza kusavimbika mune ziviso, nepo humbowo hwakasimba hunodiwa hunogona kukonzera modhi kupotsa kusadzivirirwa chaiko. Zvakawanikwa zvine chekuita nemasangano anotarisa AI-akabatsirwa kodhi ongororo kana kuchengetedza mashandiro, asi ivo havayere zvizere zvekutengesa zvekuchengetedza zvigadzirwa.
Iyo bhenji inogadzirisa kusasimba kunoshanda muAI-yakabatsirwa chengetedzo: modhi inogona kuona ine njodzi-inotaridzika pateni isinganzwisise zvinodzora zvinodzivirira kushandiswa. Batsira Net Chengetedzo inoshuma kuti kukurudzira kwakananga kwakaburitsa manyepo-akanaka mareti e41% kusvika 99%, nepo chaiyo yaitangira pa52% kusvika 71%.
Kubvunza mamodheru kuti aratidze kuti kusavimbika kunogona kushandiswa kwakaderedzwa manyepo ne17 kusvika 74 muzana, maererano neshumo, asi yakawedzera manyepo-negative mitengo kusvika pakati pe7% ne44%. AWS yakataurwa yakaderera bhawa yekugadzira yaive pazasi gumi% pamatanho ese ari maviri, uye hapana kana imwe yemagadzirirwo akaedzwa akasangana ese ari maviri maburi.
Mhedzisiro iyi haifanirwe kuverengerwa seyero yezvakakwana zvekuchengetedza zvigadzirwa. AWS yakayedzwa kamwe-kamwe-inosimudzira pamhando-yechinangwa modhi, kwete chinangwa-yakavakirwa masisitimu anoshandisa maturusi, inodzokororwa kusimbiswa, kana ajenti workflows. Saka sosi inotsigira yambiro nezve modhi-nhanho yekusagadzikana kutonga, kwete mhedziso yekuti AI kuchengetedza zvigadzirwa zvese zvinotadza.
Interactive Mechanism: Iyo Inonyatsoshanda
Ongorora ari pasi tekinoroji kuseri kwekusimudzira uku uchipindirana.
Which component of an AI application is the machine-learning model itself?
Zvekutarisa zvinotevera
Tarisa kuongororwa kwakazvimirira uchishandisa masampula akaburitswa, mhedzisiro kubva kune-turusi-kushandisa kana akawanda-nhanho chengetedzo masisitimu, uye humbowo nezvekuita munzvimbo chaidzo dzekugadzira. Iyo sosi haina kunyora mitengo, sevhisi-level yekuwana, kana kuti AWS's scoring process inowanikwa pasina zvirambidzo. Izvo zvakare hazviratidze kuti mamodheru akaedzwa anoita zvakafanana pane isina kuziviswa bhizinesi kodhi.
Vatsvakurudzi vakazvimiririra vanogona kushandisa masampuli eruzhinji uye maitiro ekuongorora pasina kudzoreredza AWS's yakashumwa mutengo wekugadzira-dhata, asi marebhurari akachengetwa uye AWS-yakasimbiswa scoring inoitirwa kudzikamisa -chaiyo optimization. Kudzokorora kunoitwa nemapoka ekunze kwaizobatsira kuyedza mhedzisiro yakashumwa uye kunaka kwemazita.
Ongororo dzenguva yemberi dzinofanira kuenzanisa modhi-yepasita imwe chete nemasisitimu anoongorora matura, kuita kodhi zvakachengeteka, kufunga pamusoro pemagadzirirwo ekutumira, uye inoda humbowo usati wapa yambiro. Iwo maedzo airatidza zvirinani kuti yakawanda sei inoshanda yekuchengetedza maturusi inovandudza pane iyo modhi-chete mhinduro.
Sosi yacho inosiya zvakakosha zvisingazivikanwe: hairatidze gumi nemaviri akaedzwa magadzirirwo emhando, taura pa-modheru mhedzisiro mune yakapihwa mameseji, mutengo wegwaro kana mamiriro ekuwana kune akavimbiswa zvibodzwa, kana kuratidza mafambisirwo ekuita kune proprietary codebases uye kurarama kuchengetedza mashandiro.