A dev set with 100 examples would not be able to detect a 0.1 percentage-point accuracy difference, such as 90.0% versus 90.1%. Compared to other machine learning problems the author has seen, a 100-example dev set is small.

Small Dev Sets Cannot Detect Tiny Accuracy Differences

Dev sets with 1,000 to 10,000 examples are common. With 10,000 examples, there is a good chance of detecting an improvement of 0.1%.

Common Dev Set Sizes for Detecting 0.1% Improvements

For mature and important applications such as advertising, web search, and product recommendations, teams may be highly motivated to find even a 0.01% improvement because it directly affects company profits. In that case, the dev set could be much larger than 10,000 examples to detect even smaller improvements.

High-Value Applications May Need Larger Dev Sets

One could test whether a change to an algorithm makes a statistically significant difference on the dev set. In practice, most teams do not do this unless they are publishing academic research papers, and the author usually does not find statistical significance tests useful for measuring interim progress.

Statistical Significance Tests on Dev Set Changes

A dev set should be large enough to detect differences between the algorithms being tried. For example, a 100-example dev set would not be able to detect a 0.1 percentage-point accuracy difference such as 90.0% versus 90.1%. Dev sets of 1,000 to 10,000 examples are common, and a dev set could be much larger than 10,000 when teams want to detect even smaller improvements.

Google

A development set, sometimes called a hold-out cross validation set, is used to tune parameters, select features, and make other decisions about the learning algorithm.

Development Set (Dev Set)

Machine Learning Yearning, a free ebook from Andrew Ng, teaches you how to structure Machine Learning projects.
https://www.deeplearning.ai/machine-learning-yearning/

Note: The content of the book is aligned with the Coursera Deeplearning.ai specialization.  https://www.deeplearning.ai/deep-learning-specialization/ 

Machine Learning Yearning (Deeplearning.ai)

Once a dev set and test set are defined, a team can try many ideas, such as different learning algorithm parameters, to see what works best. The dev and test sets allow the team to quickly see how well the algorithm is doing, and their purpose is to direct the team toward the most important changes to make to the machine learning system. Measuring each idea's performance on the dev set lets the team quickly decide whether they are heading in the right direction, and helps them identify which ideas to keep refining and which ones to discard.

Purpose of Dev and Test Sets

Dev and test sets should be chosen to reflect data one expects to receive in the future and wants to perform well on. The test set should not simply be 30% of the available data when future data is expected to differ from the training data. One should not assume the training distribution is the same as the test distribution; examples should reflect what one ultimately wants to perform well on rather than only the data available for training.

Choosing Dev and Test Sets to Reflect Future Data

Once dev and test sets are defined, a team will focus on improving dev set performance, so the dev set should reflect the task one wants to improve on the most. Having different dev and test set distributions can lead to a system that works well on the dev set but does poorly on the test set. If the dev and test sets came from the same distribution, that result would have a clear diagnosis: the system has overfit the dev set, and the obvious cure is to get more dev set data. If the dev and test sets come from different distributions, the options are less clear: the system may have overfit the dev set, the test set may be harder than the dev set, or the algorithm might be doing as well as could be expected.

Choosing Dev and Test Sets from the Same Distribution When Possible

Changing dev/test sets or evaluation metrics during a project is common. Having an initial dev/test set and metric helps the team iterate quickly. If the dev/test sets or metric no longer point the team in the right direction, change them.

Changing Dev/Test Sets or Evaluation Metrics During a Project Is Common

For a new project, choosing dev/test sets quickly gives the team a well-defined target. A practical starting point is to create initial dev/test sets and an initial metric in less than one week, even if they are imperfect. This one-week timeline does not apply to mature applications; teams working on already-mature systems may spend months acquiring better dev/test sets.

Quickly Establishing Dev/Test Sets and Metric for New Applications

Sizing the Dev Set to Detect Meaningful Accuracy Changes

What is the primary purpose of a development (dev) set in a machine learning project?

The development set is sometimes called the 'hold-out cross validation set'.

The development set is sometimes also called the _____ cross validation set.

Match each development set role to its description in Machine Learning Yearning.

Order the steps for using a dev set to choose between two model configurations.

Which of the following is NOT a stated use of the development set in Machine Learning Yearning?

The dev set and the training set refer to the same data partition in Machine Learning Yearning.

The dev set is used to tune parameters, _____ features, and make other decisions about the learning algorithm.

Match each dev set use from Machine Learning Yearning to the activity that exemplifies it.

Arrange the stages of a machine learning project that involve the dev set in the correct workflow order.

Question:
Describe the main functions of a development set in a machine learning project and provide the alternative name sometimes used for it.

Sample answer:
The development set, sometimes called the hold-out cross validation set, is primarily used to evaluate models so that the team can tune parameters, select features, and make other decisions regarding the learning algorithm.

Key points:
- Mentions the alternative name: hold-out cross validation set.
- Identifies its use for tuning parameters.
- Identifies its use for selecting features.
- Notes it is used to make other decisions regarding the learning algorithm.

Rubric:
Award full credit if the student accurately provides the alternative name and lists the three main functions of the dev set.

Explain the role and alternative terminology for the development set.

Case context:
An engineering team is building a predictive algorithm. They have a full dataset and need to decide which data columns to use as inputs (features) and what specific configuration settings to apply to their algorithm (parameters).

Question:
Which specific data partition should the team evaluate their algorithm on to make these choices, and what is another name for this partition?

Sample answer:
The team should use the development set, which is also sometimes called the hold-out cross validation set, to tune their parameters and select their features.

Key points:
- Identifies the development set as the correct partition.
- Provides the alternative name: hold-out cross validation set.
- Connects the team's need to select features and tune parameters directly to the dev set's stated purpose.

Rubric:
The answer should correctly identify the development set and its alternative name as the proper tool for the team's parameter and feature selection tasks.

Determine the correct data partition for tuning a model's parameters and features.

Question:
According to the concept text, what are the three specific actions you use a development set to perform regarding the learning algorithm?

Sample answer:
A development set is used to tune parameters, select features, and make other decisions regarding the learning algorithm.

Key points:
- Tune parameters.
- Select features.
- Make other decisions regarding the learning algorithm.

Rubric:
The response must accurately list all three actions associated with the development set.

Learn Before

Related

Learn After