Read the guide, then complete the worksheet. Submit your Idea Brief before next class.
Over the next 4 weeks, your team will design an AI system that solves a real-world problem and pitch it at the end-of-year exhibition. The best ideas get actually built.
Start with a problem, not a technology. Don't say "we want to build a neural network." Say "students struggle to review for exams — here's how AI helps."
Use this to identify which type of ML fits your idea:
| ML Type | What it does | Example use |
|---|---|---|
| Classification | Sorts inputs into categories | Spam detector, disease diagnosis |
| Regression | Predicts a numerical value | House price prediction, traffic estimation |
| Clustering | Groups similar data without labels | Customer segmentation, topic grouping |
| Recommendation | Suggests items based on behaviour | Movie recs, playlist generation |
| Generation (LLM/GenAI) | Creates new content from prompts | Chatbots, writing tools, image gen |
| Computer Vision | Understands images or video | Object detection, medical imaging |
| Reinforcement Learning | Learns by trial and reward | Game-playing agents, robotics |
Every team must address all 6 of these in their final presentation:
These are not templates — use them only if you need inspiration. The best ideas come from real problems YOU see in your daily life.
AI generates personalized quizzes from student notes using NLP
Classifies homework questions by difficulty level + suggests similar problems
Identifies struggling students from assignment patterns
Recommends NGOs to volunteers based on skills + interests (Lebanese focus)
Matches job seekers to entry-level positions based on skills + location
Recommends where donations should go based on community needs (classification)
Translates Lebanese/Gulf dialect voice notes to Modern Standard Arabic (NLP)
Flags suspicious news posts in Arabic social media (classification)
Answers school questions in Arabic 24/7 (retrieval-based NLP)
Predicts fastest commute routes in Beirut considering traffic (prediction)
Recommends parking spots based on time of day + location
Matches students going to same place for carpools
Classifies symptoms to suggest doctor specialties (not diagnose!)
Predicts when elderly patients forget medications + sends reminders
Recommends exercises based on fitness goals + preferences
Computer vision: photo of trash → sorted into recycling bins
Predicts home energy use + suggests savings (regression)
Computer vision: photo of tree → species identification + care tips
AI generates social media captions for small Lebanese businesses
Suggests items based on browsing history (recommendation)
Chatbot that handles customer questions for online shops
Pick one and ask: "What's the real problem this solves?" Then make it specific to YOUR context. E.g., not "study aid app" but "our school doesn't have a way for students to practice past exams. Our AI suggests similar questions from a database."
Before talking to your team, write 3 problem ideas on your own. Think about school, daily life, your community, Lebanon, the world.
After sharing individually, record your team's top 2 ideas below.
Complete this as a team. This is your formal submission for Week 1.