What Counts as a Rich Output?
Question: In a system that converts a recorded customer call into text, what makes the model target a rich output instead of a standard deep learning target?
Sample answer: A rich output is a result made up of many parts, such as a complete sentence, an image, or an audio clip. A standard deep learning target is usually just one numeric value.
Key points:
- Rich outputs can be long sequences or other structured media
- They are not limited to a single number
Rubric: Award full credit if the student explains that rich outputs are complex objects, while a typical deep learning target is a single value.
0
1
Tags
Data Science
Foundations of Large Language Models Course
Computing Sciences
D2L
Dive into Deep Learning @ D2L
Machine Learning
Deep Learning
Supervised Learning
Machine Learning Strategy
Machine Learning Yearning @ DeepLearning.AI
Related
In a text translation system that converts Spanish emails into English, what is the INPUT to the end-to-end model?
Direct Sentence-Pair Learning in Machine Translation
Rich-output learning in machine translation.
Match each part of a sequence-to-sequence translation task to its best description.
Order the reasoning steps for deciding whether a task is a good candidate for end-to-end learning with rich outputs.
Which statement best describes the trend in deep learning illustrated by a task with rich outputs such as speech-to-text or translation?
In end-to-end learning, rich outputs are limited to a single numeric score.
With the right (input, output) _____ pairs, a model can learn end to end even when the output is rich.
Match each example output to the correct classification as a rich output or a simple output.
Order the stages of a direct text translation system trained from paired examples.
End-to-End Translation as an Example of Rich Outputs
Classifying a Rich-Output Learning Problem
What Counts as a Rich Output?