Learn Before
Analyzing the Impact of Adding Training Data on Bias and Variance
Question: According to Andrew Ng's Machine Learning Yearning, under what conditions is adding training data viable to address high variance, and how does this action typically affect the model's bias?
Sample answer: Adding more training data is a viable solution to high variance as long as you have access to significantly more data and the computational power required to process it. This remedy is highly effective because adding training data usually reduces variance without affecting bias.
Key points:
- Adding training data requires access to significantly more data.
- Adding training data requires enough computational power to process the data.
- Adding training data typically reduces variance.
- Adding training data does not affect bias.
Rubric: The response must identify that adding data requires access to significantly more data and sufficient computational power. It must also explain that adding training data typically reduces variance while leaving bias unaffected.
0
1
Tags
Machine Learning
Deep Learning
Supervised Learning
Dive into Deep Learning @ D2L
Data Science
Machine Learning Strategy
Related
What does adding more training data primarily accomplish for a model with high variance?
True or False: Adding more training data is the simplest and most reliable way to address high variance.
Adding more training data can usually reduce _____ without affecting bias.
According to Andrew Ng, what is the simplest and most reliable way to address high variance in a machine learning model?
True or False: Adding training data typically reduces variance while also increasing bias.
Adding training data is the simplest and most _____ way to address high variance.
Match each concept to its role in Ng's guidance on adding training data to reduce variance.
Order the steps a practitioner follows when deciding to add training data as a fix for high variance.
Which pair of conditions does Andrew Ng identify as both necessary for adding training data to be a viable variance remedy?
True or False: Ng recommends adding training data for high variance only after simpler fixes like regularization have been exhausted.
Adding training data typically reduces _____ without affecting bias.
Match each descriptor to the claim Ng makes about adding training data as a variance remedy.
Order the steps for verifying that added training data successfully reduced variance without harming bias.
Analyzing the Impact of Adding Training Data on Bias and Variance
Evaluating Training Data Addition for a High Variance Speech Recognition System
Effect of Adding Training Data on Model Bias