Blog posts

2021

On providing feedback to students

3 minute read

Published:

Enye yezinto endinomdla ngazo zii-theory zokwakha nokunikezela iingxelo kubafundi bakho. Emaxesheni amaninzi xa ufundisa, uye wakhe iingxaki okanye uzithathe kwincwadi onayo, uzenze kunye nabafundi bakho eklasini. Ngoku ndandifunda kubamanga abantsi, ootishala babesenza le nto amaxesheni amaninzi (ayekhona amaxesha apho kulindeleke ingxaki uyisombulule ngowakho njengomfundi). Eyona nto ndiyikhumbulayo ngelixesha kukuba kwakulula kwaye kumnandi ukuzama ukusombulula ezo ngxaki, eklasini ninonke kunye notishala. Kwaye, le nto yayinjalo kuzo zonke izifundo, ukuqala kwisiXhosa (e.g., xa sihlalutya imibongo umzekelo) ukuya kwiMathematika.

2020

2017

2016

OntoVerbal : A Protégé plugin for verbalising ontology classes.

1 minute read

Published:

Njengokuba iimeko zibuthathaka kwizixhobo ezakhiweyo kwicala lee-ontoloji ezikwaziyo ukudilishana nelwimi zaBantu, kuyadingeka umntu ajonge izinto esele zisenziwa kwiilwimi ezifana nesiNgesi. Kukho izixhobo ezingcono kwezilwimi ngoba baninzi abantu abasebenza nazo. Isixhobo esizakusijonga namahlanje yi-OntoVerbal. I-ontoVerbal yakhiwe ngabantu abafuna kubelula ukusebenzisa i-Protégé ekufundeni inzalwano kunye nezinye izinto kwi-ontolojii nokuba umntu oziqwalaselayo akayongcali. Kumele ukuba kuyacaca ngoku kumfundi ukuba ukwakhiwa kwee-ontolojii kunezinto ezimbini ezinzima:

Cooking an ontology

2 minute read

Published:

Elinye lamaphepha endicinga limnandi lithiywe lisenzisa i-ontology yokutya ekwakheni izivakalisi xa unikwe i-RDF data [1]. Enye yezinto bendibawela ukuyazi yindlela abazikhe ngazo ezindlela zokupheka bathetha ngazo [1, p11]. Bandithumele kwiphepha loo Ribeiro et al [2], apho bathetha ngee-dialogue systems kunye nokusebenzisa kwazo ii-ontologies. Aba babhali bathi enye yendlela zokuphucula ii-dialogue systems kukwehlula kakuhle i-system kunye nolwazi malunga necala elo i-system isebenza kulo (umzekelo; elezimpilo, imozulu, etc). Le nto ingenziwa ngokusebenzisa i-intology njengento elindelwe yi-system. Eliphepha lababhali lithetha ngokwakhiwa kwe-ontology kwicala lokupheka, ontology leyo izokusetyenziswa kwi-dialogue system. I-ontology le izakuthetha ngezinto ezine; actions, food, recipes kunye nee-utensils. Inazo nee-modules ezine zokwandisa; units & measures, equivalencies kunye neentlobo zee-plates.

Real versus Template-Based Natural Language Generation

4 minute read

Published:

Namhlanje siqwalasela umsebenzi ka Deenter, Krahmer kunye no-Theune. Bathu bafuna ukubonisa ukuba iisystems ezisebenzisa iitemplates xa zikhupha imibhalo nazo ziyakwazi ukukhuphisana neesystems ezisebenzisa iigrammar xa sijonga imaintanability, usetyenziswa kolwazi lwe-lingistics kunye nomgangatho wombhalo eziwukhuphayo. Bathi zikhona iisystems zamandulo eziyibonisileyo le nto.

NLG Evaluation

2 minute read

Published:

Kwizifundo ze machine learning kukho iindlela ezimbini zokufunda okanye ukwenza umsebenzi ngee-agents. Eyokuqala yi-competition, enye yi-cooperation. Kwi-science sifunda ngokusebenzisana - umsebenzi womntu uxhomekeke kumsebenzi wonye umntu kodwa abantu bayakhuphisana kancinci. Ngamanye amazwi, idibanisa iindlela ezimbini. Kukho izifundo ezenze imikhuphiswano kuba zifuna indlela yokuphucula ulwazi. Umzekelo, kwi-robotics kukho ukhuphiswano olaziwa ngokuba yi-RoboCup. Olukhuphiswano laqalwa ngonyaka ka 1997, umsebenzi walo aluphelelanga ekuthelekiseni iincutshe, enye into kukutsala abantu abangazi nto ngezizifundo ukuba bazi banzi okanye bafuna ukufunda.

Ukwabelana, i-corpus yesiZulu

1 minute read

Published:

Ngalentsasa yanamhlanje sizokufunda umsebenzi ka Spiegler, Van der Spuy kunye no-Flach. Aba babhali bhakhe i-corpus yesiZulu, ebayithiye ngokuba nguKwabelana. Ababhali bathi isiZulu lelona lwimi linabantu abaninzi emZantsi Afrika. Bathi siyalingana ngenani nezinye iilwimi zaseNtsho (West) kodwa ziqabile izixhobo ezidilishana nesiZulu. Andisazi isizathu sokuba ababhali balinganise isiZulu nelwimi zaseNtshona ngoba xa uthetha oluhlobo wenza ingathi inani labantu abathetha ulwimi olo yeyona nto ibalulekileyo ukuba ulwimi olo lubenezixhobo, etc.

Surface realization kwi Forecast Generator (FOG)

less than 1 minute read

Published:

Ingaba i-surface realization isebenza njani kwi-FOG? Ndizive ndizibuza le nto ngenxa loyoyiko lokuba mhlawumbi ikhona into endiphosileyo ngenxa yongayifumani la ncwadi ka-Reiter no-Dale. Ndibuyele kumsebenzi ka E. Goldberg, R. Kittredge kunye no N. Driedger apho bathetha nge Forecast Production Assistant kunye ne-system efunyanwa phakathi kwayo egama lingu-FOG. Ababhali bathi ine-levels ezintlanu i-FOG kwi-surface realization. Ezi-steps zidilishana ne-abstract interlingua, syntax, ukuguqulelwa kwe-concepts kwi-sentence plans; kusetyenziswe amagama afanekileyo asuka kwi-lexicon yelwimi nganye.

Ukusebenza kwe NLG system

1 minute read

Published:

Kudala sincokola ngezinto ezininzi. Sithethe ngamaphepha kwaye savenza ingxaki esinazo ngaphepha ngaye. Ingxaki kodwa, kuyacaca ukuba khange siyijonge ncam - siyiphica-phice indlela i-NLG system eyakhiwa ngayo. Sele siqalile saveza umsebenzi we-corpus ekwakheni i-NLG system. Enye yezinto esinika uvalo kungazi ncam ukuba yakhiwa njani i-NLG system. Kulula ukuthetha nge-surface realisation kodwa kufuneka singawalibali amanye amanyathelo kwi-NLG system.

Isizathu se corpus

1 minute read

Published:

Namhlanje sibuyela emsebenzini kaDale kunye no-Reiter. Sizakucacisa izinto ebesizithatha kancinci. Eyona nto sizama ukuyiqwalasela kwakhona yindlela i-NLG system esebenza ngayo. Umbuzo wokuqala endinawo; ingaba i-corpus yeyokuthini?

Part 1 : L. Pretoius & S. Bosch

1 minute read

Published:

Phambi kokuba sithethe into ngeliphepha, kuyafuneka ukuba siliqale ngokucacisa izinto ezimbalwa. Ukuba umfundi akayazi ukuba yintoni i-morphological analyser; sizokuqala siyicacise. I-morphological analyser yi-software esetyenziswayo ekufumaneni, ekuqwalaseleni kunye nasekucaciseni amalungu olwimi; amalungu lawo zizinto ezifana neziqu zamagama, izigaba zentetho kunye nezinye izinto.

S. Sripada, N Burnett, R. Turner, J. Mastin & D. Evans

3 minute read

Published:

Kwisebe lemo-yezulu kudingwa indlela zokwenza iingxelo zemo-yezulu ngexesha elifutshane. Kudingeka iingxelo ezilungele izidingo zomntu nganye kwaye akufuneke zibe bubuvuvu ezo ngxelo. Eliphepha libonisa izizathu zokuba ababhali bacinge ukuba singakwazi ukuzifumana zonke ezizinto. Lo msebenzi uqhutywa yikomponi i-Arria NLG, UK national weather agency kunye ne-Met office.

Part 1 : Cerbah

2 minute read

Published:

Sisenengxaki yokuba ingaba xa ufuna ukuqwalasela ii-grammar yeelwimi ezimbini wenza njani. Sisazama ukubonisa umehluko phakathi kwesiZulu kunye nesiXhosa. Ingxaki ekhoyo ngoku yinto yokuba ingaba silendele ntoni ifakwekwi-program yethu ukuba sizokuba ne-program. Enye yezinto kukuba ingaba silendele impendulo ibe yinto enjani? Ingaba izakuba linani okanye izakuba yintoni?

Part 1 : Goldberg and Driedger

1 minute read

Published:

Namhlanje sizakhe sithathe umzuzu, sibuye umva ekuzameni ukwazi ukuba kubethwa abaphi kwicala lokuzama ukubonakalisa umehluko phakathi kweelwimi zesiNguni. Iphepha lanamhlanje liqwalasela ingxelo yemozulu ekhutshwa yikhompyutha. ECanada, eMelika kunye naseBhilitane kukho iiSupercomputer ezisetyenziselwa ukujonga indlela imoyezulu eqhuba ngayo. Abasasazi bemozulu bafumana ulwazi ngemo yezulu, umsebenzi wabo kukufunda olulwazi ukuze banike isishwankathelo komabona-kude kunye nakonomathotholo. Abantu abayikhatheleleyo imozulu baquka abantu nje, abafundi kunye nabanye abantu. Bakhona kodwa abantu abayikhathelele kakhulu imozulu. Aba bantu baquka amafama, abeqhubi kunye namajoni. Ulwimi olusetyenziswa kwezi zishwankathelo luxhomekeka kwintlobo yesishwankathelo luxhomekeka kwintlobo yesishwankathelo. Into ecacileyo kodwa yinto yokuba ezi zishwankathelo zisebenzisa ulwimi oluncinci kwaye alutshintshi kumaxesha athile. Bakhona abantu abadinga ulwazi oluthile kumasebe emozulu, ithi le nto lamasebe anomsebenzi omninzi ngaphezu kokushwankathela imozulu. Xa kukho iikhompyutha ezenza ezi zishwankathelo, ubomi bungaba lula. Akukho lula kodwa ukuvele uqale ukusebenzisa ezikhopyutha nokuba sele zine-program ezikwaziyo ukushwankathela. Isebe nganye libeka igama lalo esichengeni ngoba impazamo zingenza iimpazamo ezingalahlekisa abantu. Eliphepha lithetha nge-system ekuthiwa yi-Forecast Generator (FOG). Le system isebenza ngelwimi ezimbini, isiNgesi kunye nesiFrench.

Part 1 : Gxilishe, P. de Villiers & J. de Villiers

1 minute read

Published:

Eli phepha lizakucacisa ukuvumelana kwezimaphambili zesibizo kunye nesenzi. Liqwalasela indlela abantwana bayifunda njani le nto. Baqwalasela ukuba ingaba baxhomekeke kangakanani ekwazini ihlelo lesibizo. Olu lwazi luzakutyenziswa ukwenza i-models zokucacisa ukuvumelana kwezimelabizo.

Part 1 : Wang

less than 1 minute read

Published:

Eliphepha lisinika i-formal syntax framework yelwimi zabantu. Lizokucacisa ii-abstract syntax kunye nendlela ezisetyenziswa ngayo. I-linguistics licala lezifundo apho kuzanywa ukwazi ukuba iilwimi zabantu zisebenza njani. Iilwimi ziintetho kunye nemibhalo yengcinga. Ii-linguists ezininzi ziyavumelana ukuba ikhona i-universal grammar – indlela efanayo ndawu yonke yokubonisa ulwimi. Nangona kunjalo, ii-grammar zicacisa ngokupheleleyo iilwimi ezi formal. Ezi zilwimi ezifana neze-programming. Ii-grammar aziwuchani ncam ucwete xa zidilishana nelwimi zabantu.

Part 3 : Dale and Reiter

less than 1 minute read

Published:

Namhlanje sizakuhlalutya le ndaba ye-grammar engine. Ingaba kuthen ibalulekile? Sibuyela kwiphepha lika-Dale kunye no-Reiter, sikhangela ingcaciso. Ababhali bathi abantu abaninzi bayijonga njengento eyinguqulelo ye-parsing le ndaba ye-linguistic realisation. I-system yoku-parser isebenza ngoluhlobo; kusetyenziswa igrammar ax kuthathwa isivakalisi njee, sihlolwe ukuze kubonwe ukuba sithetha into kwelo lwimi luboniswa yilo-grammar. Kubalulekile ukuba sixele ngoku ukuba ezizivakalisi sithetha ngazo akunyanzelekanga ibe zezolwimi labantu. Zizivakalisi zeelwimi ze-programming, kumaxesha amaninzi. Kwinguqulelo, umsebenzi we-system ukuthatha izivakalisi ezifana nezinto ezikhutshwa zii-parser, ingqibe ngokukhupha izivakalisi ezezelwimi zabantu.

Part 2 : Twala

2 minute read

Published:

Namhlanje sizokuqhubekeka nomsebenzi kamam’ uTwala. Zizakuqala ngokujonga ii-criteria zokwazi ukuba isibizo sihlala kweliphi ihlelo. Asizokuthetha kakhulu kodwa ngalonto. Namhlanje sizokuqwalasela izimaphambili kunye nezivumelanisi. Imifanekiso uzakuyibona apha isuka isuka kumsebenzi kaTwala[4]

Part 1 : Twala

2 minute read

Published:

Umsebenzi kamama u-Edith Twala[4] uphica-phica amahlelo ezibizo kwisiZulu. Ndicinga ukuba ifani le ibizwa ngokungathi ino-H (Thwala). Isizathu salento kukuzama ukwehlula amahlelo. Ulwazi oluxelwa apha yingqokelela yezinto ekusele zisaziwa. I-chapter yokuqala izokujonga ezizinto zilandelayo:

Part 2 : Dale and Reiter

1 minute read

Published:

Zininzi indlela zokwakha i-system eyenza ezi-tasks zisithoba zee-NLG besikhe sathetha ngazo. Eyona ndlela ilula kukwakha i-module ezakwenza i-task nganye. Ii-system ezikhoyo kodwa zisebenzisa iimodule ezintathu. Ei-modules zezi :

Part 1 : Dale and Reiter

2 minute read

Published:

Namhlanje sifunda iphepha lika Reiter no-Dale lijonga ii-systems ze-NLG. Eliphepha, nangona amagama efana, lahlukile kweli lithi “Building Natural Language Generation Systems”. Kukho igama elinye elihlukileyo. Abababhali bazijonga ii-NLG systems ngengabantu abafuna ukufundisa ukuba zakhiwa njani. Bazakuthetha ngokuba zisetyenziswa nini ezinye ii-techniques ze-NLG. Banikezela ngeengcebiso malunga ne-requirements analyses kunye nezinye izinto ezenziwayo kwi-NLG. Bazakugxininisa kwi-techniques ezisele ziphucukile, le nto bayenziswa yinto yokuba ezi-techniques zingasetyenziswa ekwakheni ii-systems ezisebenzayo. Bathi bazakusinonisa ukuba ngawaphi amaphepha esingawafunda xa sidilishana nezinto ezibunzinyarha.

Part 3 : Keet Khumalo Paper

3 minute read

Published:

Namhlanje sizokuqwalasela ukuba bazakhe njani iindlela zoku-verbaliser kwisiZulu. Bakhethe izinto ezimbini abazoziqwalasela. Ezi zinto zi-quantifiers kunye ne-object properties. Zimbini izizathu zalonto. Esokuqala, ezizinto zimbini zezona sisetyenziswa kakhulu. Ithi lo nto zininzi ii-applications ezingaphuma kulomsebenzi. Isizathu sesibini sithi: sele ziliqela iilwimi ezi-controlled kwicala le NLG apho kusetyenziswa ii-Description logics ze-OWL. U-Engelbrecht kunye nogxa bakhe bazamile ukwenza isiZulu se-Science. Bayeyenziswa le nto kuba bezama ukuncedisa kwi-nursing kunye ne-healthcare. Umsebenzi lo wabo ukwazile ukuguqulela i-SNOMED CT ukuze izokwazi ukusetyenziswa ngabantu abathetha isiZulu. Umntu ofuna ulwazi olubanzi nge-SNOWMED CT angafunda i-website ye-IHTSDO[1]. Baziqwalasele zombini i-universal kunye ne-existential quantification. Baqwalasele nee-object properties apho izenzi zazo zidinga i-conjugation kunye ne-conjuction. Bathi i-disjunction ilula, abazokuyihoya ngenxa yalonto.

Part 1 : Keet Khumalo Paper

5 minute read

Published:

Ulwazi lwanamhlanje lusuka kwiphepha lika Keet kunye noKhumalo. Igama leliphepha leli lingentla. IsiZulu lulwimi lwesiNguni. IsiZulu lulwimi lwesiNguni. Iilwimi zesiNguni ziquka isiXhosa, isiNdebele kunye nesiSwati. Kuzo zonke ezilwimi, isiZulu lelona lwimi luxhaphakaleyo, sithethwa ngabantu abayi-22.7% kubo bonke abantu eMzantsi Afrika. Nangona kunjalo, isiZulu kunye nazo ezinye iilwimi esizibizileyo zinqabile kwiSoftware. Kweliphepha, ababhali bajonga ukuba ingenziwa njani iControlled natural language ukulungiselela ukubhala imithetho yeshishini okanye i-Ontology ngesiZulu. Ngamanye amazwi, kuzakuguqulelwa i-Ontology okanye ii-Business rules zisisiwa kwimibhalo yesiZulu. I-Grammaar yesiZulu (kunye nezinye iilwimi zesiNguni) iyasibonisa ukuba indlela yokusebenzisa ii-templates ayizokusebenza kulemeko. Ingxaki yinto yokuba ezilwimi zisebenzisa i-system yezibizo, iincutshe zeelwimi zithi ezilwimi zi-agglutinative kwaye zisebenzisa i-verb conjugation ene-concords kwi-class nganye yezibizo. Kweliphepha ababhali babonakalisa izinto esele bezenzile okubheka phambili.

Introducing my project

1 minute read

Published:

Ukuguqula ulwimi uluse kwelinye usebenzisa icomputer yinto enzima. Ndiyaqonda ukuba umntu ongenalwazi oluphangaleleyo ngalemeko angavele ajonge izinto ezifana noGoogle Translate acinge ukuba singabantu sele siphumelele. Indaba ezimbi yinto yokuba uGoogle Translate akasebenzi ncam.