Lesson video

Hello, my name is Mrs. Holborow and welcome to Computing.

I'm so pleased you've decided to join me for the lesson today.

In today's lesson, we're going to be exploring what the term artificial intelligence means, and we'll using an AI system to make predictions.

Welcome to today's lesson from the unit, "Machine Learning using the micro:bit.

" This lesson is called, "AI and Machine Learning Applications.

" And by the end of today's lesson, you'll be able to justify the use of artificial intelligence to solve a problem.

Shall we make a start?

We will be exploring these key words in today's lesson.

Intelligence.

Intelligence, the ability to learn and adapt to new situations.

Rule-based.

Rule-based, decisions follow fixed, predefined rules.

Data-driven.

Data-driven, decisions or predictions are based on patterns found in large amounts of data.

Look out for these keywords in today's lesson.

Today's lesson is split into three parts.

We'll start by defining the term, artificial intelligence.

We'll then move on to use an AI system to make predictions, and then we'll finish by identifying uses of a data-driven approach.

Let's make a start by defining the term, artificial intelligence.

What is intelligence?

Lucas thinks intelligence is how clever you are.

Izzy says, "Intelligence is the ability to learn and adapt to new situations.

" Both some great responses there.

Did you have anything different?

Can a machine be intelligent?

Maybe pause the video here whilst you have a think.

Lucas says, "They can adapt and learn.

" Izzy says, "But they are machines, so they are classed as artificially intelligent.

" Great point, Izzy.

Well done.

How do artificial intelligence or AI systems learn?

Let's have a look at the example of a facial recognition system.

The system uses large amounts of data such as lots of images of faces.

Software identifies and compares patterns in the data.

The system then changes how identification and comparisons are made based on feedback.

So Izzy says, "This is the ability to learn and adapt to new situations.

" Time to check your understanding.

An AI system learns by, A, remembering every data item, B, comparing patterns in data, or C, organising data into lists.

Pause the video whilst you have a think.

Did you select B?

Well done.

An AI system learns by comparing patterns in data.

An AI system can use feedback to, A, adapt the way patterns are identified, B, increase the amount of data used, or C, improve the data.

Pause the video whilst you have a think.

Did you select A?

Great work.

An AI system can use feedback to adapt the way patterns are identified.

AI systems do not improve the data they use, but they improve how they analyse it.

Okay, we're moving on to our first task of today's lesson, and I'd like you to define the term, artificial intelligence.

You could use these key words, process data, patterns, data, adapt, and feedback.

Pause the video here whilst you answer the question.

How did you get on?

Did you manage to define the term, artificial intelligence?

Let's have a look at a sample answer together.

This is only a sample answer, so don't worry if your answer differs from this.

"A system that can process data by identifying patterns in large amounts of data and adapts how this is done based on feedback.

" So you can see here we've used those words in our response.

Did you manage to use the words in your response too?

Okay, we're now moving on to the second part of today's lesson, where we're going to use an AI system to make predictions.

AI systems make predictions based on identified patterns in data.

Let's have a look at the example of TV programme recommendations.

The system uses large amounts of data such as what people watch and how long they watch it for.

Software identifies and compares patterns in the data.

The system then changes how identification and comparisons are made based on feedback.

Have you ever used a system which has recommended TV programmes for you?

How accurate was it?

Were the suggestions good?

Okay, we're now going to look at another example.

Quick Draw.

Watch the video carefully and watch the predictions the system makes at the bottom of the screen.

Here are a couple more examples.

Look closely at the predictions that are being made.

How did the system recognise the drawing or not?

Andeep says, "The system recognised the drawings by comparing the drawing with existing drawings.

" Jacob says, "And making predictions based on patterns found between existing drawings and the new drawings.

" Time to check your understanding.

What did the system use to make predictions?

Was it A, comparisons with patterns and existing drawings, B, matching the exact same existing drawing, or C, random guesses until a correct guess was made?

Pause the video whilst you have a think.

Did you select A?

Well done.

The system made predictions by comparisons with patterns in existing drawings.

True or false?

The predictions were always right first time.

Pause the video whilst you have a think.

Did you select false?

Well done.

The system has been created to make a regular prediction even if the prediction is likely to be inaccurate.

So when we first started our drawing, there wasn't much data for the system to use, so it was unlikely to be able to make an accurate prediction.

Okay, it's now time for you to have a go at Quick Draw yourself.

Go to oak.

link/quickdraw.

Have a go.

Whilst you're using Quick Draw, think about the predictions that are being made.

As a tip, it may be useful to have headphones connected for this task.

Pause the video whilst you go and complete the activity.

How did you get on?

Did you manage to do some drawings in Quick Draw?

Well done.

I'd like you to now describe how the Quick Draw system becomes more accurate as you complete more of the drawing.

Pause the video whilst you answer the question.

Here's a sample answer.

"As more parts of the drawing are given, more comparisons can be made with patterns in the data.

" I'd now like you to explain why the Quick Draw system said, "I know.

" Pause the video whilst you have a think.

Here's a sample answer.

"The Quick Draw system only said 'I know' when enough data was provided for a prediction to be made with high confidence.

" Okay, we're now moving on to the final part of today's lesson.

And you're doing a fantastic job so far, so well done.

We're now going to identify uses of a data-driven approach.

This is an example of a rule-based application.

So we have a Scratch programme with some blocks of code, and then the stage and ask sprite.

This is an example of a ChatBot.

So, if the user types in "Hello," then the sprite character will say, "Hi.

" What is a rule-based approach?

Izzy says, "A rule-based approach is where the system only responds based on programmed conditions.

" Lucas says, "The system follows predefined instructions or rules set by humans.

" It's a great couple of points there, Izzy and Lucas.

Well done.

This is another example of a rule-based application.

So here, we're applying discount in an online store.

The flow chart represents how the system works.

So if the customer is a student, then the price is going to be, price minus 10%.

If they're not, then the price is going to remain as the price.

So this means that if the customer is a student, then a 10% discount will be applied.

When are rule-based systems useful?

Andeep says, "When the problem we are trying to solve has a predictable output.

" Jacob says, "When the rules do not change often or are simple.

" Can you think of any rule-based systems that you've used?

An alternative to a rule-based application is a data-driven application.

You've already seen an example of a data-driven application in this lesson.

Quick Draw recognises drawings based on past examples.

So it's using data.

Another example of a data-driven application is the TV recommendation system.

This uses vast quantities of data collected from all users on the system to make predictions about the content the user may like.

So, for example, if you've watched a particular TV series that somebody else has, you may like similar TV programmes to what they also like.

What is a data-driven approach?

Izzy says, "A data-driven approach is where a system uses patterns in the data instead of rules.

" Andeep says, "Decisions made by the system are based on predictions.

" When are data-driven systems useful?

Andeep says, "When solving complex problems where patterns must be found.

" Jacob says, "When rules are too difficult to write or are too time-consuming.

" Time to check your understanding.

True or false?

A rule-based approach would be appropriate for a traffic light system.

Pause the video whilst you have a think.

Did you select true?

Well done.

A traffic light system follows a pre-programmed set of instructions that make use of timers and sometimes motion sensors, so it's really suitable for a rule-based approach.

Which of these is a reason for using a data-driven approach?

A, the system needs lots of data to be effective, B, the tasks are predictable, or C, the rules are too time-consuming to write.

Pause the video whilst you have a think.

Did you select C?

Well done.

One of the reasons to use a data-driven approach is that the rules may be too time-consuming to write.

Okay, we're moving on to our final set of tasks for today's lesson.

Look at these examples and decide if they are rule-based or data-driven.

So the first one is a ChatBot that always gives the same response to "How are you?

" The next one is a handwriting recognition system that improves over time.

The third one, a smart thermostat that adjusts based on past user behaviour.

And the last one, a spreadsheet that uses formulas to calculate the cost of a weekly shop.

Pause the video whilst you have a think.

How did you get on?

Did you manage to make a decision if these were rule-based or data-driven?

Let's have a look at the answers together.

So a ChatBot that always gives the same response to "How are you?

" is rule-based.

A handwriting recognition system that improves over time, is data-driven.

The key fact here is that the system is improving over time.

So with more data, the system is improving.

A smart thermostat that are just based on past user behaviour, is also data-driven, because it's looking at patterns in the previous data.

A spreadsheet that uses formulas to calculate the cost of a weekly shop, is rule-based.

Remember, if you need to go back and make any corrections, you can pause your video now.

Okay, we've come to the end of today's lesson, and you've done a fantastic job, so well done.

Let's summarise what we have learned together.

Machines can be called artificially intelligent if they can learn from and adapt to new data.

AI systems are data-driven, and make predictions based on patterns in large amounts of existing data, and can learn by adapting the way they find patterns.

Data-driven systems are different from rule-based systems, which make decisions based on fixed rules provided by humans.

Data-driven systems are useful for solving complex problems where rules are too difficult or too time-consuming to write.

I hope you've enjoyed today's lesson, and I hope I'll see you again soon.

Bye.