Explain how the training- and dev-error curves prove that more data alone is insufficient.
Question: In a concise analytical response, explain why adding training data cannot by itself reach the desired performance when training error already exceeds that level.
Sample answer: When training error is already higher than the desired performance level, adding more training data cannot move it down to that level because training error can only stay the same or get worse as data is added. The dev-error curve is usually higher than the training-error curve. Therefore, if training error cannot reach the desired level through more data alone, there is almost no way for dev error to do so.
Key points:
- Training error begins above the desired performance level.
- Training error can only stay the same or rise as data is added.
- Dev error is usually higher than training error.
- The two observations imply that dev error cannot be expected to reach the desired level through more data alone.
Rubric: A strong response identifies the initial relationship between training error and desired performance, states both source observations, and connects them logically to the conclusion that more data alone is insufficient.
0
1
Tags
Machine Learning
Deep Learning
Supervised Learning
Dive into Deep Learning @ D2L
Data Science
Machine Learning Strategy
Machine Learning Yearning @ DeepLearning.AI
Related
Why is adding more data alone insufficient when training error exceeds the desired performance level?
More data alone can bring dev error to the desired level even when training error remains above that level.
As training data increases, training error can only stay the same or _____.
Match each learning-curve element to its role in diagnosing why more data is insufficient.
Order the reasoning used to conclude that more data alone will not reach the desired performance.
Explain how the training- and dev-error curves prove that more data alone is insufficient.
Diagnose whether a team should expect more data alone to reach its desired error level.
What two observations support the conclusion that more data alone will not fix this learning curve?
Which learning-curve pattern most directly rules out more data as a sufficient solution by itself?
If training error stays unchanged as data is added, dev error can still be expected to cross below a lower desired level.