WBCHSE Applied AI

Expect a written theory paper mixing multiple-choice and very short answers with descriptive questions, plus a practical laboratory examination, project or portfolio assessment, a viva voce, and internal assessment by your school. The exact split between theory, practical and internal marks is set by the Council each session, so check the current syllabus and your school's instructions.

You need comfortable arithmetic, percentages, averages and basic probability intuition, plus the idea that a model learns patterns from examples rather than following hand-written rules. Heavy derivation is not the focus. If graphs, tables and simple statistics feel shaky, fix that early - data-handling questions lean on them more than on algebra.

Enough to write, read and debug short scripts: variables and data types, conditionals, loops, functions, lists and dictionaries, file reading, and importing libraries. Practise pandas for tables and matplotlib for charts. You should be able to explain each line of your own code aloud, because viva questions often target exactly that.

Start with AI concepts, ethics and computational thinking so you have vocabulary, then Python, then data handling and visualisation, then the machine-learning workflow, then NLP and computer vision applications. Keep coding alongside theory from week one, and start the project early enough that you can iterate on it rather than assemble it at the end.

Memorising definitions without a concrete example, writing unlabelled graphs, copying a project they cannot explain, giving one-line answers to descriptive questions that asked for reasoning, and ignoring ethics or societal-impact parts of a question. Answers that name a real application and state its limitation consistently score better than abstract definitions.

Python with Jupyter Notebook or a free cloud notebook, pandas and matplotlib for data work, and introductory scikit-learn for core models. No-code platforms such as Teachable Machine help you understand training and evaluation quickly, and public datasets let you practise cleaning and plotting. Confirm the tool list your school actually provides in the lab.

The Council publishes its compartmental and supplementary examination rules for each session, and eligibility depends on your results and your school's recommendation. There is no single fixed answer that stays true year to year, so ask your subject teacher and read the current Council notification rather than relying on what a senior told you.

It is different rather than simpler. Applied AI is broader and less mathematically deep, covering ethics, data literacy, model workflow and applications alongside programming. That breadth means more to remember but fewer long derivations. Candidates who like building small working things usually find it more rewarding; those expecting only theory often underestimate the practical work.

A small, clearly stated problem; a dataset you can describe and justify; a model or rule-based system you actually ran; honest reporting of where it failed; and clean documentation with your own screenshots, tables and conclusions. In the viva, being able to say what you would change next is worth more than claiming perfect accuracy.