Learn Before
Explain why performance alone does not define a “Small NN” in the source.
Question: In a concise analytical response, distinguish the source's definition of a small neural network from its possible performance on the same supervised learning task.
Sample answer: The source defines a small neural network structurally: it has only a small number of hidden units, layers, or parameters. It separately states that training such a network on the same supervised learning task might produce slightly better performance. Therefore, performance is a possible outcome of using the network, not the criterion that makes the network small.
Key points:
- Smallness refers to hidden units, layers, or parameters.
- Slightly better performance is presented as a possibility.
- Performance does not determine whether the network is small.
- The comparison concerns the same supervised learning task.
Rubric: A strong response clearly states the structural definition, distinguishes it from the possible performance result, and avoids claiming that slightly better performance is guaranteed.
0
1
Tags
Python Programming Language
Data Science
Machine Learning
Deep Learning
Supervised Learning
Dive into Deep Learning @ D2L
Machine Learning Strategy
Machine Learning Yearning @ DeepLearning.AI
Related
Which description best matches a small neural network in this context?
A network must have few hidden units, few layers, and few parameters simultaneously to be called small.
A small neural network has only a small number of hidden units, layers, or _____.
Match each scale indicator to its role in identifying a small neural network.
Order the reasoning steps for deciding whether a network fits the source's “Small NN” label.
Explain why performance alone does not define a “Small NN” in the source.
Decide whether a team's use of the label “Small NN” follows the source definition.
What three network characteristics does the source use to describe a small neural network?
Which claim correctly separates network scale from the reported training outcome?
A neural network with few parameters can fit the source's “Small NN” description even when performance is not provided.