Identify what a model is trying to predict
See this course’s project requirements
Evaluate the fictional box classifier and document its evidence and use boundaries in a model card.
Keep your working files in your own notebook or software. Save a private description, checklist, and work link in My Proof.
What to include
- Define the prediction task, baseline and human decision boundary.
- Document feature/label meaning and independent training/evaluation records.
- Compare two thresholds and freeze the chosen model.
- Calculate confusion counts, accuracy, precision and recall against a baseline.
- Report the separate condition audit and a targeted next evaluation step.
Review and revise
Use these criteria to check your work and choose what to improve.
- Concepts: features, labels, training and decisions remain distinct.
- Evaluation: arithmetic and denominators match the original records.
- Judgment: small samples and uneven performance are not overstated.
- Accountability: use boundaries and further evidence needs are concrete.
Want more background?
The idea in everyday life
Artificial intelligence includes many kinds of systems, and machine learning is one approach that fits patterns from data. Understanding a prediction task helps you ask better questions than whether a tool is simply intelligent. Our fictional packing lab studies scuff counts on empty practice boxes. No people are ranked, and the model will not make safety or employment decisions.
A model produces an output from inputs; people choose the task and how the output is used.
One step at a time.
Use the sample or information you have permission to use. Keep your work as you go. You can pause between steps.
- 1
Define the task: predict whether an empty practice box will receive a damaged label from a standard visual inspection. The model input is a scuff count; the output is a predicted yes or no.
- 2
Compare this with a human-written rule: "send every box with five scuffs for inspection." A threshold is a cutoff number. A learned threshold is chosen by comparing examples with known answers, even if the result looks like a simple rule.
- 3
Identify the human choices: what damaged means, which examples are collected, what errors matter, and whether anyone acts on the prediction. These choices are not eliminated by training a model.
- 4
Write a simple comparison rule, called a baseline: always predict no, or use a fixed rule without machine learning. Compare the model with this baseline before deciding whether it helps.
- 5
Write the use boundary: this is a classroom classification demonstration with invented empty boxes. All real inspection decisions remain outside scope. Save the task statement and baseline for your final model card.
See how it works.
Fictional task statement: "Given the number of visible scuffs on an empty practice box, predict the yes/no damaged label assigned by the exercise’s inspection protocol." A hand-written rule might flag five or more scuffs. A learning procedure could instead compare thresholds on labeled examples and choose the one that makes fewer training mistakes. Both produce predictions; neither observes every cause of damage or establishes that a box is suitable for a real shipment. A person still defines labels, reviews errors and decides whether the result is useful. This course investigates those relationships without calling an external AI service.
Did you cover these?
- Input and predicted label are explicit.
- A non-ML baseline is recorded.
- Human responsibility and classroom scope are named.
What to finish
A model task statement, baseline and explicit use boundary.
Watch for this
Treating a prediction as a fact or assuming every AI system learns during use.
How can I tell I understand it?
Explain the difference between model output, training and the eventual decision.
Quick knowledge check
Answer all 3 questions. Get at least 2 right and the lesson completes automatically.
You’ll get an explanation for each answer. Signed in? Your results save to your account.
Pass the check. We’ll save the progress.
Answer all 3 questions above and get at least 2 right. You can review the explanations and try again.
Go to the knowledge checkTools, templates, and lesson sources
Tools and templates
Tools open in a new tab so you can keep your place here.
Sources and further reading
Explore the references behind this lesson.
Source & access notes · checked 2026-09-10
Read MIT’s public overview of learning problems, representation, overfitting and generalization. Optional deeper study is more mathematically demanding than this original beginner course.
Access: MIT states the Open Learning Library materials are free to use without enrolling; an optional account tracks progress. This Life Starter course provides no MIT enrollment, credit, certificate or affiliation.
Reuse: Linked for reference. Examples, activities and assessments here are original; free access alone does not grant permission to reuse source text or branding.
Source & access notes · checked 2026-09-10
Opened the voluntary framework overview and human/organizational risk-management context. No compliance, fairness certification or production approval is implied.
Access: Free public reading; no account or certificate is required for this Life Starter course.
Reuse: Linked for reference. Examples, activities and assessments here are original; free access alone does not grant permission to reuse source text or branding.
Report a problem with this lesson
Tell us what went wrong. Please leave out personal information and private project contents.