Essay

Why End-to-End Models Need Labeled Structured Targets

Question: Explain what it means for an end-to-end system to predict structured outputs, give a few examples, and state what kind of training data is needed for successful learning.

Sample answer: Predicting structured outputs means the model produces results that are more complex than a single numeric value. Instead of outputting one score, the system may generate a full sentence, a sequence of labels, an object list, or another structured result. To train such a model well, you need correctly matched input-output examples for the task.

Key points:

  • Structured outputs are more complex than a single number.
  • Examples include sentences, label sequences, and object lists.
  • Training requires the right labeled input-output pairs.

Rubric: The response must explain that structured outputs are more complex than a single scalar, provide at least two valid examples, and identify the need for correctly labeled input-output pairs.

0

1

Updated 2026-08-12

Contributors are:

Who are from:

Tags

Machine Learning

Deep Learning

Supervised Learning

Dive into Deep Learning @ D2L

Data Science

Machine Learning Strategy

Machine Learning Yearning @ DeepLearning.AI