Machine learning systems do things like mistake a woman's hair for a cat or think that the name of a band from Montreal could be part of a familiar saying because

Question 1

Machine learning systems do things like mistake a woman’s hair for a cat, or think that the name of a band from Montreal could be part of a familiar saying because:

1 / 1 point

The question the QuAM was answering was too specific

QuAMs are able to generalize

They might have had limited training data

Correct! The amount of training data impacts the accuracy of the system.

Because of the learning algorithm being used

Correct! The kinds of answers a QuAM gives depend heavily on the learning algorithm used to build them.

They describe important characteristics of the data.

They are used by a QuAM to distinguish between categories

They can filter out irrelevant data

All of the above

Correct! Features capture signal from the raw data, and are used by the QuAM to perform its task. One function of features can be to filter out irrelevant information.

For some applications, the QuAM’s performance must be verified, which may mean setting up a parallel system

Correct! Running a QuAM on live data to verify performance against the existing system can be useful.

For some applications like legal rulings, explainability is important

Correct, some applications may require models to be human-interpretable.

For some applications like medical diagnosis, the bedside manner of the QuAM is more important than accuracy.

For all applications, the effects of the QuAM are far less relevant than what the system costs

For some applications there is a minimum standard of accuracy needed, and less than that standard makes the QuAM useless

Correct! You should consider what level of accuracy you need for your QuAM to be useful in operation.

For some applications, the operations of the QuAM can cause the operating environment to change, making the QuAM inaccurate or ineffective.

Correct! You need to consider how use of the QuAM will change the operating environment.

What does success look like?

What is acceptable performance?

What is ideal performance?

How are you performing by your current metrics?

Are you using the latest technology?

What is the cost of making a mistake?

How do we get perfect performance?

Are you measuring what you think you’re measuring?

It is easier to measure.

Improvements in a simple metric are likely to be real improvements to start with.

A simple proxy is easier to validate as measuring what you think it’s measuring.

All of the above

Home » Introduction to Applied Machine Learning » Machine learning systems do things like mistake a woman’s hair for a cat or think that the name of a band from Montreal could be part of a familiar saying because

Machine learning systems do things like mistake a woman’s hair for a cat or think that the name of a band from Montreal could be part of a familiar saying because

Machine Learning in the Real World Review >>> Machine learning systems do things like mistake a woman’s hair for a cat or think that the name of a band from Montreal could be part of a familiar saying because >>> Introduction to Applied Machine Learning

1.

2.

3.

4.

5.

Which of the following statements about the Machine Learning Process Lifecycle are true

Which of the following are sources of data that can be used for machine learning

What is Artificial Intelligence

What type of machine learning is the spam filter on your email platform that automatically sends suspicious emails to your spam folder