Question: In a few paragraphs, explain why it is usually better for each stage of a processing pipeline to solve a narrow, simple task, and how that choice affects the amount of training data needed.

Sample answer: A pipeline is easier to train when each stage performs a focused operation instead of a very complicated one. A narrow task usually has fewer patterns to learn, so a smaller labeled dataset can often teach that stage reliably. By contrast, a more ambitious stage may need many more examples to cover the range of situations it must handle.

This is why designers often break a system into parts that are individually learnable with the data they have. If each module is simple enough, the team can build a workable system without requiring an enormous dataset for every step. The result is a pipeline that is easier to train, test, and improve.

Key points:

• Each stage should do a limited, well-defined job
• Simpler tasks usually need fewer examples to learn well
• Data availability should influence how large or complex a stage is
• The goal is a pipeline that is practical to train and debug

Rubric: Full credit: response clearly explains that simpler stages generally require less data and connects that idea to practical pipeline design. Partial credit: response mentions either simplicity or data needs, but not the relationship between them. No credit: response is unrelated to pipeline design or training data.