Yakha Uhlelo Lokulawula Ukufunda Olunamandla E-AI Oluqeqesha Abantu Ngempela

# Isingeniso

Zibone ngeso lengqondo ubhalisela isifundo se-inthanethi, uchofoze kumaslayidi angu-40, uphase imibuzo oyisebenzise ku-Google, futhi uthole isitifiketi. Ingabe kukhona okufundile ngempela? Leli iqiniso lezindawo eziningi zokufunda eziku-inthanethi namuhla. Balandelela ukuchofoza, hhayi ukuqonda. Bakala ukuqedwa, hhayi amandla.

Izindaba ezinhle? I-Artificial intelligence yenze kwaba nokwenzeka ukwakha izinhlelo zokufunda ezivumelana nomuntu ngamunye. Amasistimu aziyo ukuthi yini osuvele uyiqonda, akhomba lapho udonsa kanzima khona, futhi akuqondise ekuphumeleleni kunokuba nje uqedele umugqa.

Kulesi sifundo, uzofunda ukuthi ulwakha kanjani uhlelo lokuphatha ukufunda olunamandla e-AI (LMS) kusukela ekuqaleni. Sizosebenzisa amathuluzi amahhala, omthombo ovulekile – akukho okubhaliselwe okubizayo kwe-API okudingekayo. Ekugcineni, uzoba nesistimu yokusebenza enezici ezine ezihlakaniphile:

• Indlela yokufunda elungisela umfundi ngamunye
• Imibuzo ekhiqizwa kabusha yi-AI
• Umfundisi wengxoxo ebukhoma enikwa amandla imodeli yolimi lwasendaweni
• Ideshibhodi elandelela ukuqhubeka kwangempela

Ungakwazi ukuhlanganisa inqolobane egcwele yephrojekthi lapha futhi ungakhohlwa ukuyinikeza a inkanyezi!

# Iyini i-LMS ene-AI?

A Uhlelo Lokuphatha Ukufunda (LMS) isofthiwe eletha, ephatha, futhi elandelela okuqukethwe kwezemfundo. Izibonelo zendabuko zihlanganisa UMoodle, Ikhanvasifuthi Ibhodi elimnyama.

I-LMS enamandla e-AI iqhubekela phambili. Esikhundleni sokukhombisa wonke umfundi okuqukethwe okufanayo ngokulandelana okufanayo, isebenzisa ubuhlakani bokwenziwa ukuze:

• Yenza kube ngokwakho ukulandelana kokufunda ngokusekelwe kulokho umfundi asekwazi kakade
• Khiqiza ukuhlola ngokushintshashintshayo kunokudonsa ebhange lemibuzo engaguquki
• Phendula imibuzo ngesiNgisi esilula ngokusebenzisa umfundisi wengxoxo
• Hlaziya idatha yokusebenza ukuze umake izindawo ezibuthakathaka futhi uphakamise izinyathelo ezilandelayo

Kucabange njengomehluko phakathi kwencwadi yokufunda kanye nomfundisi ozimele. Incwadi inikeza okuqukethwe okufanayo kuwo wonke umuntu. Umfundisi uyalungisa ngesikhathi sangempela.

# Kungani Izinkundla Zendabuko ze-LMS Ziba Mfishane

Ngaphambi kokuthi sakhe okuthile okungcono, kubalulekile ukuqonda ukuthi kungani izinkundla ezikhona zidonsa kanzima.

• Ukulethwa kokuqukethwe kosayizi owodwa: Izinkundla eziningi ze-LMS zicindezela wonke umuntu ngokuqukethwe okufanayo ngokulandelana okufanayo. Umthuthukisi ophezulu othatha isifundo sePython esaqala uchitha isikhathi emibonweni asebeyazi kakade. Oqalayo ophelele othatha izifundo ezithuthukile uyalahleka ngokushesha.
• Amabhange emibuzo amile.
  Imibuzo ebhalwe ngaphambilini yabiwa ku-inthanethi phakathi nezinsuku zokuqaliswa kwesifundo. Abafundi babamba ngekhanda izimpendulo kunokuqonda imiqondo. Ukuhlola kuba yize.
• Alukho usekelo lwesikhathi sangempela: Uma umfundi ebhajwa ngo-11pm, akekho umuntu ofundisayo ongambuza. Bangayeka noma baqhubeke ngaphandle kokuqonda okubalulekile, okuhlangana kube izinkinga ezinkulu kamuva.
• Amamethrikhi ayize ngokufunda kwangempela: Amazinga okuqeda kulula ukuwafaka umoya. Imigoqo yenqubekela-phambili nezimpawu zokuhlola zizizwa zinenzuzo kodwa azilinganisi ukuthi ulwazi ludlulisiwe ngempela yini.

Akuzona izinkinga ezincane lezi. Ngokocwaningo lwe-Research Institute of America, abafundi bagcina kuphela u-8–10% wokuqukethwe okulethwa nge-e-learning evamile. Leyo nombolo igxumela ku-25–60% ngezindlela zokufunda ezisebenzayo, eziqondene nawe. I-LMS yethu enamandla e-AI yakhelwe ukuvala lelo gebe.

# Isitaki Sobuchwepheshe Esisisebenzisayo

Sakhe le sistimu ngokuphelele ngamathuluzi omthombo ovulekile, okusho ukuthi ungawusebenzisa emshinini wakho ngezindleko ezinguziro.

// Kungani u-Ollama?

U-Ollama ikuvumela ukuthi ulande futhi usebenzise amamodeli olimi lomthombo ovulekile ngqo kukhompuyutha yakho. Awudingi i-akhawunti yefu, awukho ukhiye we-API, futhi azikho izimali zokusebenzisa. Uvele udonse imodeli bese uyibiza endaweni yokugcina ye-HTTP. Isekela amamodeli afana I-Mistral, LLAMA 3futhi Phi-3.

// Kungani iMistral 7B?

I-Mistral 7B iyimodeli encane kodwa enekhono esebenza kahle kumalaptop amaningi esimanje. Ilandela imiyalelo ngokunembile, ikhiqiza okukhiphayo okuhlanzekile kwe-JSON, futhi isingatha i-Q&A yengxoxo ngokwethembeka — lokho kanye okudingwa amamojula ethu amane.

// Kungani i-FastAPI?

FastAPI iwuhlaka lwewebhu lwesimanje lwePython olwakhelwe isivinini. Ngokwendabuko isekela ikhodi ye-asynchronous futhi Ama-WebSocketsokubalulekile ekusakazeni bukhoma izimpendulo zomfundisi esipheqululini.

# Isinyathelo 1: Izindlela Zokufunda Eziguquguqukayo

Inkinga iyayixazulula: Oqalayo kanye nonjiniyela onolwazi obhalisela isifundo esifanayo sePython akufanele balandele indlela efanayo. Imojula yokufunda eguquguqukayo ifunda iphrofayili yolwazi lomfundi ngamunye futhi yakhe ukulandelana komuntu siqu.

// Indlela Esebenza Ngayo

Uma umfundi engena emgomweni wakhe wokufunda, uhlelo luthumela isaziso kuMistral esihlanganisa:

• Amaphuzu aphezulu omfundi ngesihloko ngasinye (agcinwe emiphumeleni yemibuzo edlule)
• Uhlu lwawo wonke amamojula ezifundo atholakalayo anamazinga obunzima
• Iqoqo lemithetho: yeqa izihloko ezihlakaniphile, beka phambili izindawo ezibuthakathaka, hlonipha ukuhleleka kobunzima

I-Mistral iphendula ngohlu olu-odwe lwama-ID wemojuli — indlela yangokwezifiso yomfundi.

Isibonelo esenziwe lula kusuka main.py:

prompt = f"""
You are a curriculum expert. Return a JSON array of node IDs
in the best learning order for this learner.

Goal: {req.goal}
Mastery scores: {profile["mastery"]}
Completed modules: {profile["completed"]}
Available modules: {nodes_summary}

Rules:
- Skip completed modules
- Prioritise weak areas
- Order from easier to harder
- Return ONLY a JSON array, no explanation.
"""

Indlela ayilungisiwe. Ngaso sonke isikhathi uma umfundi eqeda imibuzo, amaphuzu akhe aphezulu abuyekezwa futhi nendlela ibalwa kabusha. Umfundi owenza kahle ngokuzumayo uthola izinto ezithuthukile ngokushesha. Umfundi odonsa kanzima uphindiselwa emuva kokuqukethwe okuyisisekelo.

// Lokho Umfundi Akubonayo

Kuthebhu ethi Indlela Yokufunda, abafundi bathayipha umgomo wabo (isb. “Funda iPython yesayensi yedatha”) bese uchofoza okuthi Khiqiza Umzila. Ngemizuzwana embalwa, kuvela ukulandelana komuntu siqu kwamamojula, ngalinye linesihloko salo, ileveli yobunzima, nezinkinobho zokugxumela ngqo kumbuzo noma kumfundisi we-AI.

# Isinyathelo sesi-2: Imibuzo Ekhiqizwe Nge-AI kanye Nokuhlola

Inkinga iyayixazulula: Amabhange emibuzo eqinile aphelelwa yisikhathi ngokushesha. Abafundi babelana ngezimpendulo, babambe ngekhanda bengaqondi, futhi basaphase. Imibuzo ekhiqizwe yi-AI ihlukile ngaso sonke isikhathi, okwenza kungenzeki ukukopela ngaphandle kokuthi ufunde ngempela.

// Indlela Esebenza Ngayo

Uma umfundi ecela imibuzo yemojuli, i-backend ibuyisa okuqukethwe kwesifundo semojuli futhi ikuthumele ku-Mistral nomyalelo oqinile wokubuyisela imibuzo ye-JSON ehlelekile.

Isibonelo esenziwe lula kusuka main.py:

prompt = f"""
Based on the following course content, generate 3 multiple-choice questions.

Topic: {node["title"]}
Content: {node["content"]}

Return ONLY valid JSON in this format:
{{
  "questions": [
    {{
      "question": "...",
      "options": ["A) ...", "B) ...", "C) ...", "D) ..."],
      "correct": "A",
      "explanation": "Short reason why this is correct."
    }}
  ]
}}
"""

Isicelo ngasinye semibuzo sikhiqiza isethi entsha yemibuzo ethathwe kumpahla yesifundo yangempela. Abafundi bathola imibuzo eyahlukene ekuzameni futhi, okuqinisa ukufunda ngokuchayeka okuhlukahlukene.

// Amaphuzu kanye Nokuhlinzeka Ngempendulo

Ngemva kokuhambisa, yonke impendulo engalungile iza nencazelo — hhayi nje ebomvu ✗. Lokhu kubalulekile. Ucwaningo ku isayensi yengqondo ibonisa ngokungaguquki ukuthi impendulo echazayo idonsela ukugcinwa okujulile kunokumane umaka izimpendulo zilungile noma zingalungile (Hattie & Timperley, 2007). Amaphuzu angu-75% noma ngaphezulu amaka imojuli njengeqediwe futhi avula izinyathelo ezilandelayo endleleni yokufunda.

# Isinyathelo sesi-3: Umfundisi we-AI Wolimi Lwemvelo

Inkinga iyayixazulula: Ukubambeka kuyisizathu sokuqala sokuthi abafundi bashiye izifundo ze-inthanethi. Ngaphandle kokuthi umuntu abuze, umzuzwana omncane wokudideka uba udonga. Umfundisi we-AI ususa lolo donga – lutholakala 24/7, ngesineke esingapheli, futhi luhlala lusekelwe kokuqukethwe kwesifundo sangempela.

// Indlela Esebenza Ngayo

Umfundisi ugijima phezu kuka-a I-WebSocket uxhumano — isiteshi esinezinhlangothi ezimbili esiphikelelayo phakathi kwesiphequluli nesingemuva. Lokhu kuvumela impendulo ye-AI ukuthi ibuyele kumsebenzisi igama negama, njengokuthayipha, kunokwenza umfundi alinde impendulo egcwele ukuze ilayishe.

Umfundisi usebenzisa indlela ebizwa ngokuthi I-Retrieval-Augmented Generation (RAG). Ngaphambi kokuphendula, idonsela okuqukethwe kwesifundo esifanele ekwazisweni njengomongo. Lokhu kusekela izimpendulo zikaMistral ezintweni zesifundo sangempela kunolwazi olujwayelekile, kunciphisa ubungozi bezimpendulo ezingalungile noma ezingabalulekile.

Isakhiwo sokwaziswa esenziwe lula:

prompt = f"""
You are a concise, helpful programming tutor.
Answer based on the context below. If the answer is not in the
context, say so and give a general answer.

Course Context: {node_content}

Previous conversation:
{conversation_history}

Learner: {user_message}
Tutor:
"""

Umlando wengxoxo ufakiwe kuyo yonke imilayezo, ngakho umfundisi ukhumbula okwashiwo ekuqaleni kweseshini efanayo, okwenza ingxoxo izwakale ingokwemvelo kunokuba iphindaphinde.

// Lokho Umfundi Akubonayo

Kuthebhu ye-AI Tutor, abafundi babona isixhumi esibonakalayo sengxoxo esijwayelekile. Bathayipha umbuzo, cindezela u-Enter, bese ubuka ukusakazwa kwempendulo ngethokheni. Uma bezulazula besuka kumojula ethile, umfundisi useyazi ngokuqukethwe kwaleyo mojula njengomongo.

# Isinyathelo sesi-4: Ukulandelela Inqubekelaphambili Nezibalo

Inkinga iyayixazulula: Amadeshibhodi amaningi akubonisa ibha yephesenti egcwala njengoba uchofoza okuqukethwe. Lokho akusona isilinganiso sokufunda; kuyisilinganiso sokuchofoza. Ideshibhodi yethu ilandelela ukuphumelela ngesihloko, okwakhiwe ngokusebenza kwemibuzo yangempela ngokuhamba kwesikhathi.

// Indlela Esebenza Ngayo

Konke ukuthunyelwa kwemibuzo kubangela izinto ezimbili:

1. Isibuyekezo sesikolo esihle sisebenzisa i-Exponential Moving Average (EMA)

New mastery = 30% recent score + 70% historical mastery
new_mastery = 0.3 * quiz_score + 0.7 * current_mastery

I I-Exponential Moving Average inikeza isisindo esengeziwe ekusebenzeni kwakamuva ngenkathi isafaka umlando. Umfundi obelokhu edonsa kanzima kodwa osanda kuphumelela uzobona amaphuzu akhe aphezulu ekhuphuka, kodwa anganyuki ngaso leso sikhathi ngomphumela owodwa omuhle. Lokhu kwenza imethrikhi ibe qotho.

2. Inqubekelaphambili umcimbi ungene

Sonke isenzo – kusukela ekuqaleni kwemojuli kuya ekuhambiseni imibuzo, ukuphasisa noma ukwehluleka – kufakwe isitembu sesikhathi. Lokhu kudala irekhodi eligcwele lomsebenzi wokufunda onika ideshibhodi amandla.

// Lokho Umfundi Akubonayo

Ithebhu yedeshibhodi ibonisa:

• Amamojula aqediwe kunani eliphelele elitholakalayo
• Izinga lokuqeda njengephesenti
• Ukuphumelela okumaphakathi kuzo zonke izihloko ezifundiwe
• Amabha okuphatha isihloko — okuluhlaza okotshani okunemibala (okuqinile), inhlaka (iyathuthuka), noma okubomvu (okubuthakathaka)
• Igridi yesimo semojuli: ukubuka konke okubonakalayo kokuthi yimaphi amamojula enziwayo futhi asele

Lokhu kunikeza abafundi isithombe sangempela sokuthi bami kuphi, hhayi nje ukuthi sebehambe ibanga elingakanani.

# Indlela Wonke Amamojula Amane Asebenza Ngayo Ndawonye

Imojula ngayinye iwusizo iyodwa, kodwa ndawonye bakha iluphu yempendulo eqhubekayo.

Iluphu yempendulo yomfundi

Le loop isho ukuthi isistimu ayilokothi ime. Iphendula indlela umuntu ngamunye asebenza ngayo, hhayi nje ukuthi uchofozile “Olandelayo.”

Izakhiwo zasendaweni ezigcwele — awekho amafu, awekho okhiye be-API

# Isiphetho

Ukwakha i-LMS enamandla e-AI akudingi isabelomali esikhulu noma ithimba lesayensi yedatha. Nge-Ollama, FastAPI, kanye ne-React, ungakha isistimu evumelana ngempela nabafundi — leyo eyenza ukuhlola okusha, ephendula imibuzo ngesikhathi sangempela, futhi ilandelele ukubamba kahle kwangempela kunokuba nje ukuqedele.

Okwenza le ndlela ibe namandla akusona isici esisodwa. I-loop yempendulo. Uhlelo luhlakanipha ngomfundi ngamunye ngayo yonke imibuzo ethunyelwe, yonke imibuzo ebuzwayo, kanye nemojula ngayinye eqediwe.

Izinkundla zendabuko ze-LMS zilandelela ukuchofoza. Lokhu kulandela ukufunda.

Iphrojekthi egcwele – ehlanganisa yonke imizila yasemuva, izingxenye ze-React, nemiyalo yokusetha – iyatholakala ku-GitHub. Isondeze futhi ufunde i-README ukuze uyiqalise endaweni.

Shithu Olumide ungunjiniyela wesofthiwe nombhali wezobuchwepheshe othanda ukusebenzisa ubuchwepheshe obuphambili ekwenzeni izindaba ezithokozisayo, oneso elibukhali lemininingwane kanye nekhono lokwenza imiqondo eyinkimbinkimbi ibe lula. Ungathola futhi i-Shittu Twitter.