IB COMPUTER SCIENCE | A4 ASSESSMENT

MACHINE LEARNING: END OF TOPIC REVIEW

Check learning fundamentals, hardware, labelled HL methods and ethical implications through scored retrieval, applied tasks and developed written responses.

BEFORE YOU BEGIN

Use two or more sessions if needed: first the automatic checks and applied tasks, then the written responses. Allow time to show working and develop discussion answers. Common SL/HL content is A4.1 and A4.4; A4.2 and A4.3 are HL-only.

Sections A and B award one point for each fully correct automatic item, with no partial credit within an item. Section C has separately suggested content marks. These are original practice questions, not an official IB paper, grade boundary or prediction.

Choose your level below. Write calculations, diagrams and Section C answers in your exercise book before revealing guidance. Answers and selections last while the page remains open.

Automatic total: 48 points. Suggested written total: 84 marks, assessed separately.

A | QUICK CHECKS

Answer without notes. The common questions are A1–A8; A9–A24 are HL-only. In Section B, B1–B8 are common and B9–B24 are HL-only.

A1 | Which learning descriptions are accurate?

Select all correct answers. One point requires the complete correct set.

A2 | What task estimates a measured delivery duration?
A3 | Which distinguish training and inference?

Select all correct answers. One point requires the complete correct set.

A4 | Which hardware statements are accurate?

Select all correct answers. One point requires the complete correct set.

A5 | Which concerns remain after an explicit sensitive column is removed?

Select all correct answers. One point requires the complete correct set.

A6 | Which make human oversight meaningful?

Select all correct answers. One point requires the complete correct set.

A7 | Which distinguish AR and VR?

Select all correct answers. One point requires the complete correct set.

A8 | Which justify renewed ethical review?

Select all correct answers. One point requires the complete correct set.

A9 | HL: Which are appropriate cleaning principles?

Select all correct answers. One point requires the complete correct set.

A10 | HL: A method ranks features using an association criterion without comparing fitted final-model subsets. Which strategy is it?
A11 | HL: Which operations reduce input dimensionality?

Select all correct answers. One point requires the complete correct set.

A12 | HL: Which interpret a fitted line accurately?

Select all correct answers. One point requires the complete correct set.

A13 | HL: Which describe K-NN classification?

Select all correct answers. One point requires the complete correct set.

A14 | HL: Which tuning practices preserve evaluation?

Select all correct answers. One point requires the complete correct set.

A15 | HL: Which describe K-means?

Select all correct answers. One point requires the complete correct set.

A16 | HL: Which association-rule statements are accurate?

Select all correct answers. One point requires the complete correct set.

A17 | HL: Which describe reinforcement learning?

Select all correct answers. One point requires the complete correct set.

A18 | HL: Which describe a genetic route search?

Select all correct answers. One point requires the complete correct set.

A19 | HL: Which describe an ANN computation?

Select all correct answers. One point requires the complete correct set.

A20 | HL: Which distinguish CNN settings and learned parameters?

Select all correct answers. One point requires the complete correct set.

A21 | HL: Which justify task-specific model comparison?

Select all correct answers. One point requires the complete correct set.

A22 | HL: Which identify leakage risks?

Select all correct answers. One point requires the complete correct set.

A23 | HL: Which distinguish approaches?

Select all correct answers. One point requires the complete correct set.

A24 | HL: Which are cautious evaluation claims?

Select all correct answers. One point requires the complete correct set.

B | APPLY YOUR KNOWLEDGE

Each applied item gives its own assumptions. Show calculations, a sketch or a short explanation in your book before entering a final answer. Each item is self-contained. Evaluate its stated dataset and rules rather than carrying changed data from another question.

B1 | A shop groups purchasing records without known segment labels. Which task best matches the aim?
B2 | Which configurations are plausible starting points under stated workload checks?

Select all correct answers. One point requires the complete correct set.

B3 | Ten million raw model parameters use four bytes each. How many raw bytes do the parameters occupy, excluding all other state?
B4 | A platform removes a quoted abusive phrase in a report while missing contextually abusive messages. Which responses need consideration?

Select all correct answers. One point requires the complete correct set.

B5 | A hiring system excludes an applicant based on a wrong record. Which measures directly support accountability and correction?

Select all correct answers. One point requires the complete correct set.

B6 | In a class of 30, 18 have compatible devices. Six different students without devices receive loans. How many still lack compatible equipment under this simple simultaneous-use assumption?
B7 | An AR translation update starts identifying people. Which conclusions are justified?

Select all correct answers. One point requires the complete correct set.

B8 | Which statements about long-term quantum-related privacy are careful?

Select all correct answers. One point requires the complete correct set.

B9 | HL: Training minimum=4 and maximum=8. What is the min–max transform of x=5?
B10 | HL: A selector repeatedly fits a chosen model with candidate feature subsets and compares validation error. Which strategy is used?
B11 | HL: With five bins per input dimension and six dimensions, how many possible cells are in the simplified coverage grid?
B12 | HL: For ŷ=2.4x +3.8, predict y at x=4.5.
B13 | HL: Observed y=8 and predicted ŷ=9. Which describes the residual y −ŷ?
B14 | HL: With k=3, selected neighbour labels are A, B, A under equal voting. What is predicted?
B15 | HL: A tree tests Fragile? If Yes predict Special; otherwise test mass ≤2kg, with Yes→Standard and No→Special. For non-fragile mass exactly 2kg, what is predicted?
B16 | HL: Positive means defective. TP 16, FP 8, FN 4, TN 72. What is recall as a percentage?
B17 | HL: For TP 16, FP 8, FN 4, TN 72, what is F1 as a percentage rounded to one decimal place?
B18 | HL: Cluster members are (2,1),(8,2),(8,3),(9,2). What is the updated centroid?
B19 | HL: In 100 baskets X occurs 20 times, Y30 times and both 12 times. What is lift for X→Y?
B20 | HL: An action gives −1 and the following terminal action gives +8. What is the undiscounted cumulative reward?
B21 | HL: Prefix crossover copies AB from ABCDEF, then scans parent D A F C B E and appends unused stops. Which child results?
B22 | HL: x₁=2,x₂=1,w₁=0.6,w₂=−0.4,b=−0.3. What is z before activation?
B23 | HL: A 5×5 input uses a 3×3 kernel, stride 2, no padding and dilation 1. How large is each output feature map?
B24 | HL: A tree has recall 75% and memory 4 MiB; K-NN 85% and 32 MiB; ANN 90% and 12 MiB. With recall ≥85% and memory ≤8 MiB, which is eligible?

CHECK SECTIONS A AND B

Unanswered items receive zero. Each checkbox item requires all correct choices and no incorrect extras. Written marks are separate. Changing an answer or level requires resubmission.

C | WRITTEN COMMUNICATION AND SUBJECT KNOWLEDGE

Write in your exercise book before revealing the guidance. The table gives practical guidance based on the IB command terms supplied in the guide; it is not the Cambridge command-word table.

In particular, IB compare focuses on similarities, while distinguish asks for differences. If you need to address both, the question can request each separately. All main commands below are highlighted.

IB command term What to provide in this review Writing approach
Calculate A numerical result with relevant working stages Show the calculation and units
Describe A detailed account Give features or an ordered process, not only a name
Explain A detailed account including reasons or causes Link ideas with because, therefore, resulting in or leading to
Compare Similarities, referring to both items throughout Both…; each…; in both approaches…
Compare and contrast Similarities and differences, referring to both throughout Use both for similarities and whereas for differences
Discuss A considered, balanced review of relevant arguments or factors Develop strengths and limits in context and reach a supported judgement
Distinguish Clear differences between the concepts Whereas…; unlike…; one uses distances whereas the other follows learned tests…
Construct Information displayed in diagrammatic or logical form Supply the requested table, map or connected diagram
Deduce A conclusion reached from the supplied information Use the map or data as evidence
Justify Valid reasons or evidence for a choice Connect the choice to the stated requirement
Evaluate An appraisal weighing strengths and limitations Balance relevant factors and reach a supported judgement

A connective earns no automatic mark by itself: it must express an accurate relationship. For example, “a large unscaled feature can dominate distance, therefore K-NN may select misleading neighbours under that representation” explains a mechanism and consequence. Suggested marks reward subject content, not extra unofficial grammar marks.

C1 | LEARNING SETTINGS

6 suggested marks | COMMON SL/HL

Compare and contrast supervised and unsupervised learning. Distinguish reinforcement learning from both using a robot-navigation example.

C2 | JUSTIFY HARDWARE

6 suggested marks | COMMON SL/HL

A battery-powered vision robot must react locally while a research team trains a large model remotely. Justify suitable hardware and infrastructure considerations for both. Include processing, memory, storage and one scalability or deployment constraint.

C3 | SCHOOL SUPPORT ETHICS

6 suggested marks | COMMON SL/HL

Discuss using historical referrals to prioritise students for support. Include an intended benefit, a mechanism of unfairness, privacy and meaningful human review.

C4 | REASSESS AN EXPANDED TECHNOLOGY

6 suggested marks | COMMON SL/HL

AR glasses initially translate signs but an update identifies people and displays inferred profiles. Discuss whether the original ethical judgement should cover the update. Include rights, privacy, equity and reassessment.

C5 | ONLINE COMMUNICATION

6 suggested marks | COMMON SL/HL

Discuss automatic moderation of harassment while allowing anonymous or pseudonymous participation. Consider safety, expression, privacy and appeal.

C6 | FUTURE CAPABILITY AND DAILY LIFE

6 suggested marks | COMMON SL/HL

Discuss how quantum-related future capabilities and compulsory VR activities can create privacy or equity responsibilities before a serious incident occurs.

C7 | HL: PREPARE THE DATA

6 suggested marks | HL ONLY

A route dataset includes repeated imports, mixed metres/kilometres, missing distances and a verified rare long journey. Explain a preparation policy and why learned transformations must be fitted inside training.

C8 | HL: REGRESSION AND FIT

6 suggested marks | HL ONLY

For a fitted line ŷ=2.4x +3.8, observed inputs cover 1–5 km. SSE=4.4 and SST=62 on the training sample. Calculate the prediction at 4.5 km and R², then explain slope, intercept and why a prediction at 7 km needs caution.

C9 | HL: TWO CLASSIFICATION MECHANISMS

6 suggested marks | HL ONLY

Compare and contrast K-NN and a decision tree for categorical predictions. Include prediction steps, the role of labelled examples and one representation or complexity limitation for each.

C10 | HL: METRICS AND SELECTION

6 suggested marks | HL ONLY

Defective is positive. A validation matrix has TP 16,FP 8,FN 4,TN 72. Calculate precision and recall. Explain why model selection must also consider error costs, hard constraints and independent testing.

C11 | HL: GROUPS AND ASSOCIATIONS

6 suggested marks | HL ONLY

Distinguish customer clustering from association-rule mining. A rule has 100 transactions, antecedent count 20, consequent count 30 and joint count 12. Calculate support, confidence and lift, then state an interpretation limit.

C12 | HL: REWARD LEARNING AND GENETIC SEARCH

6 suggested marks | HL ONLY

A robot can finish for +2 immediately or take −1 then +8. A separate optimiser searches full tours. Explain exploration/exploitation for the robot and describe the population-search mechanism for the optimiser.

C13 | HL: ANN STRUCTURE

6 suggested marks | HL ONLY

Construct labelled sketches of a single perceptron and a 2–2–1 MLP. Explain why nonlinear hidden activations matter and distinguish forward calculation from training.

C14 | HL: CNN HIERARCHIES

6 suggested marks | HL ONLY

Describe a CNN classification pipeline and how it adaptively learns spatial feature hierarchies. Include the effects of kernel size, stride, activation and loss.

CHECK THE QUALITY OF YOUR WRITING

Did you answer the command? For explain, check the reason or cause; for compare, address both items through similarities; for compare and contrast, include differences too; for discuss, balance relevant factors in context; for distinguish, make the differences explicit; for evaluate, weigh strengths and limitations and support your conclusion.

For construct, check labels, connections and completeness. For calculate, show enough working to follow your method and label units. Select two weak responses and rewrite them after revising. Keep the written total separate from your automatic result.

YOUR NEXT REVISION STEP

Submit the automatic sections for targeted revision links.

Record your own written marks using the suggested points. Choose two weaknesses and attempt new examples after revising. Use the A4 revision page for worked content and a personal flashcard list, or return to the IB course index.