Question: In a few paragraphs, explain why the source recommends that each pipeline component be a relatively simple function, and how this relates to the amount of data needed to learn that component.

Sample answer: The source explains that when building a pipeline, each component should ideally be a relatively 'simple' function. This matters because simple functions can be learned effectively from only a modest amount of data, whereas more complex functions would typically require larger datasets to learn well. By keeping each pipeline component's task simple, a team increases the chance that the available training data will be sufficient to learn that component accurately, making the overall pipeline more practical to build and train.

Key points:

• Each pipeline component should be a relatively simple function
• Simple functions can be learned from only a modest amount of data
• This guides the choice of what components to include in a pipeline
• The goal is to make each component learnable given the data available

Rubric: Full credit: response accurately states that simplicity of a component reduces the data needed to learn it, and connects this to the goal of building a workable pipeline with modest data. Partial credit: response mentions simplicity or data requirements but not both, or the connection is unclear. No credit: response is unsupported by the source or off-topic.