Ukuvumelana (Alignment)
Alignment
Inqubo yokuqinisekisa ukuthi izinhloso, imiphumela, kanye nokuziphatha kohlelo lwe-AI kuvumelana nezinhloso nezindinganiso zabantu. Lokhu kubaluleke kakhulu ezinhlelweni ezithuthukile ezingase zithuthukise izindlela ezingahlosiwe ngokuqondile.
Isibonelo: Ukuqinisekisa ukuthi ingxoxo ye-chatbot yezempilo yengqondo ayinikezi izinyathelo eziyingozi noma ngabe yimiphi imiyalezo.
Isikhombimsebenzisi Sesicelo Sokuhlela (API) (Application Programming Interface (API))
Application Programming Interface (API)
Iqoqo lemithetho nezinqubo ezichaziwe ezivumela izinhlelo ezihlukahlukene zesoftware ukuthi zixhumane futhi zishintshane ngemininingwane.
Isibonelo: Ukusebenzisa i-API ye-OpenAI ukuthumela umyalezo nokuthola impendulo eyenziwe imodeli yolimi kuhlelo lwakho lokusebenza.
Ubuhlakani Bokwenziwa Obujwayelekile (AGI) (Artificial General Intelligence (AGI))
Artificial General Intelligence (AGI)
Ifomu elithakazelisayo le-AI elingenza noma yimuphi umsebenzi wokucabanga umuntu angawenza. Liyakwazi ukufunda ezindaweni ezahlukene.
Isibonelo: Uhlelo lwe-AGI lungafunda ukudala umculo, lwenze imisebenzi yobudokotela, futhi ludlule isivivinyo sefilosofi ngaphandle kokuhlelwa okukhethekile.
Ubuhlakani Bokwenziwa (AI) (Artificial Intelligence (AI))
Artificial Intelligence (AI)
Ukulingisa ubuhlakani bomuntu kumishini ehlelwe ukucabanga, ukucabanga, nokuziphatha ngokuzimela.
Isibonelo: I-AI isebenza abasizi bomuntu njengoSiri nezinhlelo zokushayela ezizimele njenge-Tesla Autopilot.
Izimiso Zokuziphatha Ze-AI (AI Ethics)
AI Ethics
Isiyalo esikhathazekile ngezinkinga zokuziphatha zokuthuthukiswa nokusetshenziswa kwe-AI, okuhlanganisa ukulunga, ubumfihlo, ukuziphendula, nokungakhethi.
Isibonelo: Ukudala imihlahlandlela yokuvimbela izinhlelo zokuqasha ukuthi zingakhethi ngokobulili noma ubuhlanga.
Ubuhlakani Obuqiniswe Ngokwengeziwe (Augmented Intelligence)
Augmented Intelligence
Imodeli yokubambisana lapho i-AI ihambisana futhi yengeza ubuhlakani bomuntu kunokuba iyibuyisele.
Isibonelo: Izinsiza ze-AI zokuthwebula izithombe ezikhombisa izinto ezingajwayelekile kodokotela, abenza isinqumo sokugcina.
I-ejenti Ezimele Yodwa (Autonomous Agent)
Autonomous Agent
Uhlelo lwe-AI olukwazi ukwenza izinqumo zalo futhi luthathe izinyathelo ukuze lufeze izinhloso zalo ngaphandle kokuphazamiseka komuntu.
Isibonelo: Iloli lokulethwa elizihambelayo elizulazula emigwaqweni yedolobha futhi ligweme izithiyo ngokuzimela.
Ukubuyela Emumva (Backpropagation)
Backpropagation
Indlela yokuqeqesha amanethiwekhi ezinzwa ngokuvuselela izisindo ngokubuyela emuva kusuka kumiphumela kuya kumalayeri wokufaka, kunciphisa amaphutha okubikezela.
Isibonelo: Kusetshenziswa ekuqeqesheni izithombe ezihlukanisa izithombe ukunciphisa izinga lamaphutha ekubhaleni izinombolo ezibhalwe ngesandla.
Ukuthambekela (Ukuthambekela Okubalwayo) (Bias (Algorithmic Bias))
Bias (Algorithmic Bias)
Ukuchema okungahlosiwe nokuhlelekile emiphumeleni ye-AI ngenxa yedatha yokuqeqesha engalingani noma engameleki.
Isibonelo: Uhlelo lokubona ubuso olubona kabi abantu abanemibala esikhunjeni ngenxa yokungameleli kwedatha yokuqeqesha.
Idatha Enkulu (Big Data)
Big Data
Amaqoqo amakhulu kakhulu eminingi adinga amathuluzi akhethekile ukugcina, ukuhlaziya, nokukhipha inani, avame ukusetshenziselwa ukuqeqesha amamodeli e-AI.
Isibonelo: Ukusebenzisa izigidi zezinto ezisebenzisanayo ukuqeqesha izinjini zokuncoma ezindaweni ze-e-commerce.
Imidwebo Eyimfihlo (Black Box Model)
Black Box Model
Uhlobo lwemodeli ye-AI noma yokufunda ngomshini okungekho kalula ukuchaza ingqondo yayo eyimfihlo, okwenza kube nzima ukuqonda ukuthi izinqumo zenziwa kanjani.
Isibonelo: Inethiwekhi enzulu yezinzwa esetshenziselwa ukuvumela izikweletu kodwa ayinikezi incazelo ecacile yokuthi kungani umfakisicelo owodwa wamukelwa kanti omunye wanqatshwa.
Ukubala Okuhlakaniphile (Cognitive Computing)
Cognitive Computing
Izinhlelo ze-AI eziklanyelwe ukulingisa izinqubo zokucabanga zomuntu, njengokucabanga nokufunda, kusetshenziswa izindlela ezifana ne-NLP nokubona amaphethini.
Isibonelo: Uhlelo lokubala oluhlakaniphile olusiza abachwepheshe bezomthetho ukuhlaziya umthetho wecala nokubikezela imiphumela.
Umbono Wekhompyutha (Computer Vision)
Computer Vision
Isizinda sobuhlakani bokwenziwa esivumela amakhompyutha ukuthi achaze futhi acubungule idatha ebonakalayo njengezithombe nevidiyo.
Isibonelo: Izinhlelo zokubona ubuso ezibona abantu kumavidiyo okuqapha kusetshenziswa umbono wekhompyutha.
Umbhalo Omkhulu (Corpus)
Corpus
Iqoqo elikhulu lemibhalo ebhaliwe noma ekhulunyiwe esetshenziselwa ukuqeqesha izimodeli zolimi.
Isibonelo: I-Common Crawl dataset iyimbhalo yewebhu yomphakathi esetshenziselwa ukuqeqesha izimodeli zolimi ezinkulu njenge-GPT.
Ukudonswa Kwemininingwane (Data Drift)
Data Drift
Isenzakalo lapho idatha yokufaka ishintsha ngokuhamba kwesikhathi, okubangela ukwehla kokusebenza kwemodeli.
Isibonelo: Imodeli yokugcinwa okubikezelayo yemishini yezimboni iba nokunemba okuncane njengoba kungeniswa ubuchwepheshe obusha bezinzwa.
Ukulebula Idatha (Data Labelling)
Data Labelling
Inqubo yokufaka amalebula emininingwaneni ngamalebula noma amagama-ukhiye ukuze ilungele ukufunda okubhekiwe.
Isibonelo: Ukulebula izithombe eziningi zomdlavuza njengezingenawo umdlavuza noma ezine-malignant ukuqeqesha imodeli yokuthola umdlavuza.
Ukumba Idatha (Data Mining)
Data Mining
Inqubo yokuthola amaphethini acacile, ukuhlobana, nokungajwayelekile kumazinga amakhulu eminingi.
Isibonelo: Abathengisi basebenzisa ukumba idatha ukuthola ukuthi abantu abathenga ama-nappy bavame ukuthenga nebhiya.
Ukufunda Okujulile (Deep Learning)
Deep Learning
Isizinda sokufunda ngomshini esisebenzisa amanethiwekhi ezinzwa anezigaba eziningi ukumodela amaphethini ayinkimbinkimbi kudatha.
Isibonelo: Ukufunda okujulile kusetshenziswa ezimodeleni zolimi njenge-GPT-4 kanye nezimodeli zokukhiqiza izithombe njenge-Stable Diffusion.
Amamodeli Okudiliza (Diffusion Models)
Diffusion Models
Uhlobo lwezimodeli zokukhiqiza ezifunda ukukhiqiza idatha ngokuguqula kancane umsindo ongahlelekile ube yimiphumela ehlelekile.
Isibonelo: I-Stable Diffusion idala izithombe ezingokoqobo kusuka kumiyalezo yombhalo isebenzisa izindlela zokudiliza.
Ukufakwa (Embedding)
Embedding
Ukumelelwa kwedatha nge-vector, okusetshenziselwa ukubamba incazelo yezincazelo zamagama, izithombe, noma imisho.
Isibonelo: Ku-NLP, igama elithi 'ibhange' lingaba nokufakwa okufanayo negama elithi 'imali' kodwa lihlukile kwelithi 'umngcele womfula' kuye ngokombhalo.
I-Epoch (Epoch)
Epoch
Ukuphindwa okugcwele phezu kwedatha yokuqeqesha yonke phakathi nenqubo yokuqeqesha imodeli yokufunda ngomshini.
Isibonelo: Uma idatha inezibonelo eziyi-1,000 futhi imodeli ibona zonke kanye, lokho kuyi-epoch eyodwa.
I-AI Ezinzile (Ethical AI)
Ethical AI
Ifilosofi yokuklama nokusebenzisa eqinisekisa ukuthi ubuchwepheshe be-AI busebenza ngokucacile, ngokulinganayo, futhi ngokuhambisana nezindinganiso zomphakathi.
Isibonelo: Ithuluzi lokuqasha i-AI elihlanganisa izinsiza zokuhlola ukuthambekela ukuze kuvinjwe ukucwaswa kwabazobhapathizwa abavela emiphakathini encane.
Uhlelo Lwezingcweti (Expert System)
Expert System
Uhlelo lwe-AI olulingisa amakhono okwenza izinqumo engcweti yomuntu endaweni ethile kusetshenziswa imithetho nokucabanga.
Isibonelo: Uhlelo lwezingcweti olusetshenziswa ezolimo ukuncoma ukwelashwa kwezitshalo ngokusekelwe kudatha yenhlabathi nomlando wezinambuzane.
I-AI Echazwayo (XAI) (Explainable AI (XAI))
Explainable AI (XAI)
Izinhlelo ze-AI eziklanyelwe ukwenza izinqubo zazo zangaphakathi kanye nezinqumo ziqondakale kubantu, zikhuphule ukwethenjwa nokuziphendula.
Isibonelo: I-AI yokuxilonga yezokwelapha enikeza isiphakamiso kuphela kodwa futhi ichaza ukuthi yiziphi izimpawu eziholele kuleso siphetho.
Ukufunda Okunemibono Embalwa (Few-shot Learning)
Few-shot Learning
Indlela yokufunda ngomshini lapho imodeli iqeqeshwa noma ilungiswa kusetshenziswa izibonelo ezimbalwa ezinelebuli kuphela.
Isibonelo: Ukwenza ngokwezifiso i-LLM ukuthi ibhale ama-imeyili asemthethweni ngemva kokuyibonisa izibonelo eziyi-10 kuphela.
Ukulungisa (Fine-tuning)
Fine-tuning
Inqubo yokuthatha imodeli eqeqeshwe ngaphambili bese uyiyoqeqesha kabanzi kudatha entsha, encane ukuze uyenze ibe nekhono emsebenzini othile.
Isibonelo: Ukulungisa i-LLM ejwayelekile njenge-GPT kumadokhumenti angaphakathi asemthethweni ukuze kwakhiwe umsizi wokubhala wezomthetho.
Imidwebo Eyisisekelo (Foundation Model)
Foundation Model
Imodeli enkulu eqeqeshwe kudatha ehlukahlukene futhi ebanzi engalungiswa imisebenzi eminingi.
Isibonelo: I-GPT-4 ne-PaLM 2 yizimodeli eziyisisekelo ezikwazi ukufingqa, ukubuzwa, ukuhumusha, nokunye okuningi.
Izinto Ezihlukahlukene (Fuzzy Logic)
Fuzzy Logic
Uhlobo lokucabanga olubhekana namazinga angalungile kunokucabanga okuyiqiniso/okungelona iqiniso (okubili), okusetshenziselwa ukucabanga ngaphansi kokungaqiniseki.
Isibonelo: Kusetshenziswa ezinhlelweni zokulawula isimo sezulu ukulungisa izinga lokushisa ngokusekelwe kumazinga angalungile afana 'nokushisa kancane' noma 'okubandayo kakhulu'.
Inethiwekhi Yokukhiqiza Eyinqaba (GAN) (Generative Adversarial Network (GAN))
Generative Adversarial Network (GAN)
Isakhiwo semodeli yokukhiqiza lapho izinhlelo ezimbili - umkhiqizi nomhlukanisi - ziqhudelana ukuthuthukisa ikhwalithi yomphumela.
Isibonelo: Ama-GAN asetshenziselwa ukudala amavidiyo omgunyathi noma ukukhiqiza izithombe zomkhiqizo ezingokoqobo kusuka kumidwebo.
I-AI Yokukhiqiza (Generative AI)
Generative AI
Isigaba sobuhlakani bokwenziwa esingakha okuqukethwe okusha - njengombhalo, izithombe, umculo, noma ividiyo - kusuka kudatha yokuqeqesha.
Isibonelo: I-ChatGPT ikhiqiza izindatshana zamabhulogi noma i-Midjourney idala ubuciko bedijithali kusuka kumiyalezo yombhalo.
Isihumushi Esikhulu Esilungiselelwe Ngaphambili (GPT) (Generative Pre-trained Transformer (GPT))
Generative Pre-trained Transformer (GPT)
Uhlobo lwezimodeli zolimi ezinkulu ezakhiwe yi-OpenAI ezisebenzisa isakhiwo sesihumushi futhi iqeqeshwe ngaphambili kudatha yombhalo omkhulu ukuze yenze imisebenzi eyahlukene yolimi.
Isibonelo: I-GPT-4 iyakwazi ukubhala izindatshana, ukuhumusha izilimi, nokufingqa amadokhumenti ngemiyalezo encane.
Umshini Wofuzo (Genetic Algorithm)
Genetic Algorithm
Indlela yokwenza okungcono kakhulu ephefumulelwe ukukhethwa kwemvelo lapho izixazululo ziguquka ngokuhamba kwesikhathi ngokusebenzisa ukuguquka, ukuhlanganisa, nokukhetha.
Isibonelo: Kusetshenziselwa ukuklama izakhiwo zenethiwekhi ezinzwa ezisebenzayo ngokulinganisa ukusinda kwezifanele.
Ukuhalalisela (Hallucination)
Hallucination
Ukuphuma kokuqukethwe okuzwakalayo kodwa okungalungile ngokweqiniso noma okungenangqondo yi-AI model.
Isibonelo: Imodeli yolimi yenza isicatshana esingekho noma inikeza amaqiniso omlando angalungile.
Isiqondiso (Heuristic)
Heuristic
Indlela esebenzayo yokuxazulula izinkinga engaqinisekisi isixazululo esiphelele kodwa esanele izinhloso ezisheshayo.
Isibonelo: Ukusebenzisa umthetho wesithupha ukuze kulinganiselwe isikhathi sokulethwa ohlelweni lwe-AI lwezinto zokuhambisa.
I-Hyperparameter (Hyperparameter)
Hyperparameter
Inani lokumisa elibekwe ngaphambi kokuqeqesha imodeli yokufunda ngomshini, njengesilinganiso sokufunda noma inani lezendlalelo.
Isibonelo: Ukulungisa usayizi weqoqo kusuka ku-32 kuya ku-128 ukuthuthukisa isivinini sokuqeqesha nokusebenza kwemodeli.
Ukuboniswa (Inference)
Inference
Inqubo yokusebenzisa imodeli yokufunda ngomshini eqeqeshwe ukwenza izibikezelo noma ukukhiqiza imiphumela kusuka kudatha yokufaka entsha.
Isibonelo: Ukusebenzisa imodeli ye-GPT elungisiwe ukudala ama-imeyili eqembu losizo lwamakhasimende.
Ukutholwa Kwesifiso (Intent Detection)
Intent Detection
Umsebenzi ekuqondeni ulimi olungokwemvelo lapho uhlelo luthola injongo yomsebenzisi noma umgomo emyalezweni.
Isibonelo: Ku-chatbot, ukuqaphela ukuthi 'Ngifuna ukubhuka indiza' njengesifiso sokubhuka uhambo.
Inethiwekhi Yezinto (IoT) (Internet of Things (IoT))
Internet of Things (IoT)
Inethiwekhi yezinto ezisebenza ngokuxhumana ezihlanganisa izinzwa, isoftware, nezinye izobuchwepheshe ukuqoqa nokushintshana ngemininingwane.
Isibonelo: Ama-thermostat ahlakaniphile namafriji abika idatha yokusetshenziswa futhi alungise izilungiselelo kusetshenziswa izibalo ze-AI.
Ukuchazeka (Interpretability)
Interpretability
Izinga lapho umuntu angakuqonda khona izindlela zangaphakathi zemodeli yokufunda ngomshini kanye nenqubo yayo yokwenza izinqumo.
Isibonelo: Isihlahla esinqumayo sichazeka kakhulu kunenethiwekhi enzulu yezinzwa ngoba izinqumo zayo ziyalandeleka.
I-Jupyter Notebook (Jupyter Notebook)
Jupyter Notebook
Indawo yokubala esebenza ngokubambisana evulekile eyivumela abasebenzisi ukuthi babhale ikhodi, babonise imiphumela, futhi babhale ukuhlaziywa esikhombimsebenzisi esisodwa.
Isibonelo: Ososayensi bedatha basebenzisa i-Jupyter Notebooks ukwenza amamodeli okufunda ngomshini nokwabelana ngemiphumela.
I-K-Nearest Neighbours (KNN) (K-Nearest Neighbours (KNN))
K-Nearest Neighbours (KNN)
I-algorithm elula, engaphathelene yokufunda ngomshini esetshenziselwa ukuhlukanisa nokubala. Yenza izinqumo ngokusekelwe ezibonelweni zokuqeqesha eziseduze esikhundleni sezici.
Isibonelo: Ukuhlukanisa isithelo esisha njengosawolintshi noma ipheya, i-KNN ihlola ukuthi yiziphi izithelo ezinelebuli eziseduze ngesimo nombala.
Igrafu Yolwazi (Knowledge Graph)
Knowledge Graph
Isakhiwo sedatha esisebenzisa ama-node nemiphetho ukumelela nokugcina izincazelo ezixhunyiwe zezinto kanye nobudlelwano bazo.
Isibonelo: Iphaneli yolwazi ye-Google inikwa amandla yi-knowledge graph exhumanisa izinto ezifana nabantu, izindawo, nezehlakalo.
Ukulungiswa Kwemodeli Yokufunda Ulimi (LLMO) (Language Learning Model Optimisation (LLMO))
Language Learning Model Optimisation (LLMO)
Izindlela ezisetshenziselwa ukuthuthukisa ukusebenza, ukusebenza kahle, noma ukulungiswa kwezimodeli zolimi ezinkulu zemisebenzi noma izindawo ezithile.
Isibonelo: Ukusebenzisa ukubalwa nokulungiswa kwemiyalezo ukulungisa i-LLM ukusetshenziswa kwezinga lebhizinisi.
Imimodeli Ulimi Olukhulu (LLM) (Large Language Model (LLM))
Large Language Model (LLM)
Uhlobo lwemodeli yokufunda ejulile eqeqeshwe ngamanani amakhulu edatha yombhalo ekwazi ukukhiqiza, ukuqonda, nokucabanga ngolimi lwabantu.
Isibonelo: I-ChatGPT ne-Claude yizimimodeli zolimi ezinkulu eziqeqeshwe ukusiza ekubhaleni, ukuhlela, nokuphendula imibuzo.
Isikhala Esifihlekile (Latent Space)
Latent Space
Ukumelelwa okuyinkimbinkimbi okunamandla amaningi lapho izinto zokufaka ezifanayo ziqoqwa eduze, kusetshenziselwa izimodeli zokukhiqiza nokufakwa.
Isibonelo: Ekukhiqizeni izithombe, ukulungisa isikhala esifihlekile kungashintsha izici njengokukhanya noma umuzwa.
Isilinganiso Sokufunda (Learning Rate)
Learning Rate
I-hyperparameter eyinhloko ekuqeqesheni elawula ukuthi izisindo zemodeli zilungiswa kangakanani maqondana ne-gradient yelahleko.
Isibonelo: Isilinganiso sokufunda esiphakeme singaholela ekudluleleni ngaphezulu kwama-minima, kanti isilinganiso esiphansi kakhulu siyalibazisa inqubekela phambili yokuqeqesha.
Ukufunda Ngomshini (ML) (Machine Learning (ML))
Machine Learning (ML)
Isikhundla se-AI esivumela izinhlelo ukuthi zifunde kusuka kudatha futhi zithuthukise ukusebenza ngaphandle kokuhlelwa ngokuqondile.
Isibonelo: Izihlungi ze-Spam zisebenzisa ukufunda ngomshini ukuhlukanisa ama-imeyili njenge-spam noma cha ngokusekelwe ezibonelweni ezedlule.
Ukudonswa Kwemodeli (Model Drift)
Model Drift
Isenzakalo lapho ukunemba kwemodeli kwehlisa ngokuhamba kwesikhathi ngenxa yezinguquko kudatha noma imvelo.
Isibonelo: Imodeli yokuthola inkohlakalo iba nokunemba okuncane njengoba izindlela zenkohlakalo ziguquka.
Ukuqeqeshwa Kwemodeli (Model Training)
Model Training
Inqubo yokudlulisela idatha kumodeli yokufunda ngomshini nokulungisa amaparamitha ayo ukunciphisa iphutha.
Isibonelo: Ukuqeqesha injini yokuncoma ngomlando wokuthenga wamakhasimende ukuze kunconywe imikhiqizo emisha.
I-AI Ye-Multimodal (Multimodal AI)
Multimodal AI
Izinhlelo ze-AI ezikwazi ukucubungula nokuhlanganisa izinhlobo eziningi zemibhalo njengombhalo, izithombe, umsindo, nevidiyo.
Isibonelo: Imodeli efana ne-GPT-4 Vision ekwazi ukufunda umbhalo nokuchaza izithombe ngesikhathi esifanayo.
Ukucubungula Ulimi Olungokwemvelo (NLP) (Natural Language Processing (NLP))
Natural Language Processing (NLP)
Isizinda se-AI esigxile ekusebenzelaneni phakathi kwamakhompyutha nezilimi zabantu (ezemvelo). Ivumela imishini ukuthi ifunde, iqonde, futhi iphendule ngolimi lwabantu.
Isibonelo: I-NLP isetshenziswa kubasizi bezwi, izinhlelo zokuhumusha ulimi, kanye nama-chatbots.
Inethiwekhi Yezinzwa (Neural Network)
Neural Network
Imodeli yokufunda ngomshini ephefumulelwe isakhiwo sobuchopho bomuntu, equkethe izendlalelo zamalungu axhunyiwe (ama-neuron).
Isibonelo: Amanethiwekhi ezinzwa angemuva kwezimodeli zokufunda ezijulile ezisetshenziswa ekubhaleni izithombe nasekuqashelweni kwezwi.
Umsindo (Noise)
Noise
Ulwazi olungahlelekile noma olungabalulekile kudatha olungafihla amaphethini acacile futhi luthinte kabi ukusebenza kwemodeli.
Isibonelo: Amaphutha ezinzwa noma okufakiwe kwedatha okuneziphutha kungabhekwa njengomsindo.
I-Ontology (Ontology)
Ontology
Isakhiwo esihlelekile esihlukanisa futhi sichaze ubudlelwano phakathi kwemibono endaweni ethile, esetshenziselwa izinhlelo ze-AI ezinencazelo.
Isibonelo: I-ontology kwezokwelapha ingachaza ukuthi izimpawu zihlobana kanjani nezifo nokwelashwa.
Ukweqisa (Overfitting)
Overfitting
Iphutha lokumodela lapho imodeli yokufunda ngomshini ibamba umsindo kudatha yokuqeqesha futhi yenza kabi kudatha entsha.
Isibonelo: Imodeli egcina izimpendulo zokuqeqesha kodwa ayikwazi ukubhekana nedatha engabonwanga ihlelwe ngokweqisa.
Izibalo Zokubikezela (Predictive Analytics)
Predictive Analytics
Ukusetshenziswa kwedatha, izibalo, ne-AI ukuthola ukwenzeka kwemiphumela yesikhathi esizayo ngokusekelwe kudatha yomlando.
Isibonelo: Abathengisi basebenzisa izibalo zokubikezela ukubikezela isidingo semikhiqizo ethile.
Ukuqeqeshwa Kwangaphambili (Pre-training)
Pre-training
Inqubo yokuqeqesha imodeli ekuqaleni kudatha enkulu, ejwayelekile ngaphambi kokuyilungisa imisebenzi ethile.
Isibonelo: Amamodeli e-GPT aqeqeshwa ngaphambili kumakhorasi amakhulu ngaphambi kokwenziwa ngokwezifiso ukuba abe ama-chatbot osizo lwamakhasimende.
Ubunjiniyela Bezimiso (Prompt Engineering)
Prompt Engineering
Ubuciko nesayensi yokwakha imiyalezo esebenzayo ukuze kuqondiswe imiphumela yezimodeli zolimi ezinkulu.
Isibonelo: Ukwengeza imiyalelo yohlelo efana 'Ngaphendula njengomfundisi ohloniphayo' kuyisibonelo sobunjiniyela bezimiso.
Ukubalwa (Quantisation)
Quantisation
Indlela yokunciphisa imodeli eyanciphisa inani lamabhithi asetshenziselwa ukumelela izisindo nokubalwa, kuthuthukisa ukusebenza kahle.
Isibonelo: Ukubala imodeli kusuka ku-32-bit kuya ku-8-bit kuthuthukisa ukusebenza kumadivayisi eselula.
Ukubala Kwe-Quantum (Quantum Computing)
Quantum Computing
Ipharamitha entsha yokubala esekelwe ku-quantum mechanics, enikeza amandla okucubungula aphindaphindayo.
Isibonelo: Ukubala kwe-Quantum kungase kusheshe ukuqeqeshwa kwe-AI ngaphezu kwemikhawulo yakudala.
Injini Yokucabanga (Reasoning Engine)
Reasoning Engine
Uhlelo ku-AI olukhipha iziphetho ezinengqondo kusuka eqoqweni leziqinisekiso noma idatha kusetshenziswa imithetho noma izibalo zokuboniswa.
Isibonelo: Ithuluzi lokuxilonga le-AI lisebenzisa injini yokucabanga ukukhipha izimo zezokwelapha ezingenzeka ngokusekelwe ezimpawini.
Ukufunda Ngokuqinisa (RL) (Reinforcement Learning (RL))
Reinforcement Learning (RL)
Isizinda sokufunda ngomshini lapho ama-ejenti afunda ngokusebenzisana nemvelo yawo ukuze akhulise imivuzo eqongekayo.
Isibonelo: Inhloli efunda ukuhamba ngokuzama nokuhluleka kusetshenziswa izindlela ze-RL.
Ukufunda Ngokuqinisa Ngempendulo Yomuntu (RLHF) (Reinforcement Learning with Human Feedback (RLHF))
Reinforcement Learning with Human Feedback (RLHF)
Indlela yokufunda lapho izintandokazi zabantu ziqondisa isignali yemvuzo ye-AI, evame ukusetshenziselwa ukulungisa izimodeli zolimi.
Isibonelo: I-ChatGPT yaqeqeshwa nge-RLHF ukuze ikhiqize izimpendulo ezisiza kakhulu futhi eziphephile.
Ukukhiqiza Okungenelela Okutholakele (RAG) (Retrieval-Augmented Generation (RAG))
Retrieval-Augmented Generation (RAG)
Indlela ehlanganisa ukuthola ulwazi nokukhiqiza, lapho i-LLM ithola amadokhumenti afanele ukuthuthukisa impendulo yayo.
Isibonelo: Umsizi we-AI uthola futhi ucaphuna imininingwane yomkhiqizo ngenkathi ekhiqiza impendulo kumbuzo wezobuchwepheshe.
Ukufunda Okuzenzakalelayo (Self-Supervised Learning)
Self-Supervised Learning
Indlela yokuqeqesha lapho imodeli ifunda amaphethini ngokukhiqiza amalebula ayo kusuka kudatha eluhlaza, kunciphisa ukuncika kudatha enezincazelo zabantu.
Isibonelo: I-BERT iqeqeshwa ngokufunda okuzenzakalelayo ngokubikezela amagama alahlekile embhalweni.
Ukusesha Okunenjongo (Semantic Search)
Semantic Search
Indlela yokusesha eqonda isifiso somsebenzisi nencazelo yombhalo, hhayi nje ukufanisa amagama-ukhiye.
Isibonelo: Ukusesha 'indlela yokulungisa ithephu evuzayo' kubuyisa imihlahlandlela noma ngabe igama elithi 'ithephu evuzayo' alikho kudokhumenti.
Ukuhlaziywa Komzwelo (Sentiment Analysis)
Sentiment Analysis
Inqubo yokuthola imizwa, imibono, noma izimo zengqondo embhalweni, ngokuvamile ihlukanisa njengenhle, engalungile, noma engathathi hlangothi.
Isibonelo: Ukuhlaziya ama-tweet ukuze kulinganiswe ukusabela komphakathi kumkhiqizo omusha.
Okungahlelekile (Stochastic)
Stochastic
Kuhlanganisa ukuziphatha okungahlelekile noma okungokwenzeka, okuvame ukusetshenziselwa i-AI yokukhiqiza kanye nezibalo zokwenza okungcono kakhulu.
Isibonelo: Umphumela we-GPT-4 uyahlukahluka ngomsebenzi ofanayo ngenxa yenqubo yayo yokukhipha engahlelekile.
I-AI Eqinile (Strong AI)
Strong AI
Kuye futhi kubizwe nge-Artificial General Intelligence (AGI), kubhekiselwa kumishini enekhono lokucabanga elilingana nelomuntu kuzo zonke izindawo.
Isibonelo: I-AI yesikhathi esizayo engakwazi ukubhala ngokuzimela amanoveli, ukuhlela amadolobha, nokuxazulula izinkinga zokuziphatha ngokulinganayo.
I-Super Artificial Intelligence (SAI) (Super Artificial Intelligence (SAI))
Super Artificial Intelligence (SAI)
I-AI ethakazelisayo edlula kakhulu ubuhlakani bomuntu kuzo zonke izici - ukucabanga, ubuciko, ubuhlakani bezenhlalo, njll.
Isibonelo: I-SAI ingakwazi ukuthuthukisa izifundo ezintsha zesayensi nefilosofi ngokuzimela.
Ukufunda Okunakekelwe (Supervised Learning)
Supervised Learning
Indlela yokufunda ngomshini lapho izimodeli ziqeqeshwa ngedatha enezincazelo ukufunda ukufaniswa kokufaka-okuphumayo.
Isibonelo: Ukufundisa imodeli ukuhlukanisa ama-imeyili njenge-spam noma cha kusetshenziswa izibonelo zomlando.
Idatha Eyenziwe Ngokwenziwa (Synthetic Data)
Synthetic Data
Idatha eyenziwe ngokwenziwa efanisana nedatha yomhlaba wangempela, evame ukusetshenziselwa ukuqeqesha lapho idatha yangempela ingavamile noma ibucayi.
Isibonelo: Ukudala izithombe zezokwelapha ezenziwe ngokwenziwa ukuqeqesha izimodeli zokuxilonga ngaphandle kokwephula ubumfihlo besiguli.
I-Token (Token)
Token
Iyunithi yombhalo ecubungulwa yi-LLMs - ngokuvamile igama noma ingxenye yegama.
Isibonelo: Umusho othi 'Sawubona mhlaba!' uhlukaniswe izigaba ezi-3: 'Sawubona', 'mhlaba', kanye '!'.
Ukuthokozisa (Tokenisation)
Tokenisation
Inqubo yokuhlukanisa umbhalo ube izigaba ukuze kucutshungulwe imodeli.
Isibonelo: Ku-NLP, 'I-ChatGPT inhle' iba yi-['Chat', 'G', 'PT', 'iyinhle'].
Ukufunda Okudlulisiwe (Transfer Learning)
Transfer Learning
Ukusebenzisa ulwazi oluvela komunye umsebenzi ukuthuthukisa ukufunda komunye umsebenzi ohlobene, kunciphisa isikhathi sokuqeqesha nezidingo zedatha.
Isibonelo: Ukulungisa imodeli eqeqeshwe embhalweni wesiNgisi ukwenza ukuhlaziywa komzwelo kolunye ulimi.
Isihumushi (Transformer)
Transformer
Isakhiwo senethiwekhi esisebenzisa izindlela zokunaka ukumodela idatha elandelanayo, esetshenziswa kabanzi kuma-LLM.
Isibonelo: I-BERT, i-GPT, kanye ne-T5 zonke ziyizimodeli ezisekelwe kusihumushi.
Ukwehluleka (Underfitting)
Underfitting
Lapho imodeli ilula kakhulu ukuthi ingabamba amaphethini kudatha yokuqeqesha, okuholela ekusebenzeni okubi.
Isibonelo: Imodeli elinganiselayo ezama ukubikezela ukuhlukaniswa kwezithombe eziyinkimbinkimbi ingase yehluleke.
Ukufunda Okungazenzisi (Unsupervised Learning)
Unsupervised Learning
Indlela yokufunda lapho izimodeli zithola amaphethini noma amaqoqo kudatha engezona ezinelebuli.
Isibonelo: Ukuqoqa amakhasimende ngokusekelwe ekuziphatheni kokuthenga ngaphandle kwamalebula achaziwe.
Isifiso Somsebenzisi (User Intent)
User Intent
Umgomo noma injongo ngemuva kokufuna noma ukusebenzisana komsebenzisi.
Isibonelo: Umsebenzisi othayipha 'indlela yokubhaka ikhekhe' cishe ufuna iresiphi.
Isethi Yokuvumela (Validation Set)
Validation Set
Ingxenye yedatha esetshenziselwa ukuhlola ukusebenza kwemodeli phakathi nokuqeqesha nokulungisa ama-hyperparameters.
Isibonelo: Kusetshenziselwa ukuthola ukweqisa ngaphambi kokuhlolwa kokugcina.
Idathabhesi Ye-Vector (Vector Database)
Vector Database
Idathabhesi eyenzelwe ukugcina nokusesha ama-vector embeddings asetshenziswa emisebenzini ye-AI njengokusesha okufanayo kanye ne-RAG.
Isibonelo: I-Pinecone ne-Weaviate yizidathabhesi ze-vector zokugcina imibhalo noma ukufakwa kwezithombe.
Ukufakwa Kwama-Vector (Vector Embedding)
Vector Embedding
Ukumelelwa kwedatha okunombolo okugcina incazelo nezinhlobonhlobo zobudlelwano esikhundleni se-vector.
Isibonelo: Amagama athi 'inkosi' kanye 'nendlovukazi' anokufakwa okufanayo okunomehluko wobulili obucayi.
Umsizi We-Virtual (Virtual Assistant)
Virtual Assistant
I-ejenti yesoftware enikwe amandla yi-AI esiza abasebenzisi bafeze imisebenzi ngokuxoxisana noma ngemiyalezo yezwi.
Isibonelo: I-Siri, i-Alexa, kanye ne-Google Assistant yizisizi ezidumile ze-virtual.
Ukuqashelwa Kwezwi (Voice Recognition)
Voice Recognition
Ubuchwepheshe obuchaza futhi buguqula ulimi olukhulunywayo lube umbhalo noma isenzo.
Isibonelo: Ukuthayipha ngezwi kanye nemiyalezo yezwi kuncike ezinhlelweni zokuqashelwa kwezwi.
I-AI Ebuhlakani (Weak AI)
Weak AI
Izinhlelo ze-AI eziklanyelwe ukwenza umsebenzi owodwa, othize ngaphandle kobuhlakani obujwayelekile.
Isibonelo: I-AI yokudlala ishachisi engakwazi ukuqonda ulimi noma ukushayela imoto iyisibonelo se-AI ebuhlakani.
Ukuklwebha I-Web (Web Scraping)
Web Scraping
Ukuqoqwa okuzenzakalelayo kolwazi kusuka kumawebhusayithi, okuvame ukusetshenziselwa ukuqoqa idatha yokuqeqesha noma ukuqapha okuqukethwe.
Isibonelo: Ukuklwebha izinhlu zezindlu ukuze kuqeqeshwe imodeli yokulinganisa impahla.
Isisindo (Weight)
Weight
Ipharamitha kumanethiwekhi ezinzwa enquma amandla omthelela ilungu elilodwa linaye komunye.
Isibonelo: Izisindo zilungiswa ngesikhathi sokuqeqesha ukunciphisa iphutha lemodyuli.
Ukukhuluma (Whisper)
Whisper
Imodeli yokukhuluma-kuya-umbhalo eyakhiwe yi-OpenAI ekwazi ukubhala umsindo ngezilimi eziningi.
Isibonelo: Ukukhuluma kungabhala izinkulumo namaphodikhasti ngokunemba okuphezulu.
YAML (YAML)
YAML
Ifomethi efundeka kalula yokuqondanisa idatha, evame ukusetshenziselwa amafayela okumisa emisebenzini yokufunda ngomshini.
Isibonelo: Ukuchaza amapharamitha emodeli kufayela le-YAML ukuze kuqeqeshwe ku-PyTorch.
Ukufunda Okunolwazi Oluncane (Zero-shot Learning)
Zero-shot Learning
Amandla emodeli okwenza imisebenzi ayizange iqeqeshwe ngayo ngokukhethekile ngokusebenzisa ulwazi olujwayelekile.
Isibonelo: Imodeli ephendula imibuzo yezomthetho noma ngabe ayizange iqeqeshwe ngokukhethekile ngedatha yezomthetho.
Zettabyte (Zettabyte)
Zettabyte
Iyunithi yedatha yedijithali elingana nama-byte ayi-sextillion (10^21), evame ukusetshenziselwa ukuchaza isilinganiso sedatha ye-inthanethi.
Isibonelo: Umkhawulo womhlaba we-inthanethi udlule i-zettabyte eyodwa ngonyaka ngo-2016.
