Speaking · Free lesson
ML English: training data or prediction input?
A teammate sees actual delivery durations in old records and asks whether every new shipment needs that value before the model can predict it.
Professor Scott MendozaThe short answer
Say “The training records include actual delivery durations. For a new prediction, the model uses distance and package weight; it predicts the duration.” In this fictional task, the outcome being predicted is not supplied as a prediction input.
Name the stage and the field’s role.
Training is the stage in which this supervised model learns from examples with known outcomes. Inference means using the trained model to make a prediction. For this example, distance and package weight are inputs; actual duration is the known outcome in historical records and the value predicted for a new shipment. Use during training and for a new prediction to make the contrast explicit.
Compare the examples
During training, the records include the actual delivery duration.
For a new prediction, we provide distance and package weight.
The model predicts the delivery duration.
Required during training and required for a new prediction are not interchangeable phrases.
Build the meaning step by step
Stage
During training
Known outcome
actual delivery duration
Different stage
For a new prediction, distance and package weight are the inputs.
Try it now
Separate the roles
Read all answers and explanations
Which phrase names the stage that uses the trained model for a prediction?
During inferenceInference is using the trained model for a prediction in this context.
In the supplied shipment example, which pair is provided for a new prediction?
Distance and package weightThe supplied inputs are distance and weight; duration is the predicted value.
Choose the statement that preserves the distinction.
Historical records include actual durations; the new prediction estimates a duration.A historical outcome and the value being predicted have different roles.
Put it into practice
Explain the data roles
Use the supplied example: distance and weight go in; a duration prediction comes out. Historical records include actual durations for training.
- Alex
The old records contain actual delivery durations. Do we need that value for each new prediction?
See a model answer
- Alex
The old records contain actual delivery durations. Do we need that value for each new prediction?
- You
The actual duration is the known outcome during training. It is not an input for a new prediction in this example.
- Alex
Then what do we provide for a new shipment?
- You
We provide distance and package weight, and the model predicts the delivery duration.
The contrast resolves the teammate’s question without claiming all models work identically.
A little more clarity
Common questions
A few useful details before you move on.
Does every ML model use this exact setup?
No. This lesson uses one simplified supervised example. Name the actual task and fields when explaining another system.
In this example, duration is the value being predicted.
What does label mean in this context?
It is the known target value in a labeled training example, not a shipping sticker. Here it is the actual delivery duration.
The label in each historical example is its actual delivery duration.
Take it into your day
Separate training outcomes, prediction inputs and the predicted output.
English for Careers · Coming Soon
Explain ML work to your team.
English for Machine Learning Engineers: Career Simulation is coming soon. The planned course follows connected predictive-ML workplace situations. Sign up on the course page for its launch news.
Get launch updatesMachine Learning Engineers · Career Simulation
Your next step · Available on Udemy
Strengthen your English foundations.
The available English Grammar and English Speaking Master Course develops broader grammar, sentence-building and communication foundations. Clear contrasts and sentence structure help explain technical roles.
Explore the course English Master Course: English Grammar and English Speaking