What's the estimated number of IoT devices in use by the end of 2020?

Smartphones are IoT devices

IoT devices need to be connected to the Internet at all time

IoT devices are always secure

AI can be run on low powered IoT devices

The T in IoT stands for:

IoT devices gather information from the world around them using:

IoT devices draw more power than the average desktop or laptop computer

The three steps in a CPU instruction cycle are:

What operating system do Raspberry Pis run?

IoT devices typically run faster and have more memory than the average desktop or laptop computer

An LED is a sensor:

Sensors are used to:

Actuators are used to:

Digital sensors send data as:

What digital signal is sent when a button is pressed?

You can control analog actuators from a digital device using pulse-width modulation

IoT devices always need to be connected to the Internet to work:

IoT devices always communicate over HTTP, the same as web apps and other web APIs:

IoT devices rely on the cloud for all their decision making:

Data gathered from sensors and sent to the cloud is called:

What should happen to a command if the IoT device is offline:

MQTT, or Message Queueing Telemetry Transport, has message queues:

IoT devices can be used to support agriculture

Plant needs include: (pick the best answer)

Sensors are too expensive for farmers in developed nations to use:

Plant growth is dependent on temperature

The temperatures to consider for plant growth are:

Growing Degree Days are calculated using which formula:

IoT devices can be used to detect ambient properties like soil moisture

Which of these can cause problems for plant growth (pick one)?

All sensors are provided pre-calibrated for standard units:

When measuring soil moisture, one difference between resistive and capacitive moisture sensors is:

The SPI protocol supports:

The I2C protocol supports:

IoT devices are powerful enough to control water pumps

Actuators can be used to control power to additional devices

Sensors detect changes from actuators immediately

A relay is what type of switch:

A relay allows:

Actuators should always respond instantly to sensor readings:

A public MQTT broker is fine to use for a commercial IoT project

Cloud computing allows you to:

There are multiple cloud vendors with data centers in many different countries across 6 continents

To request control of an actuator and get a response from an IoT device, app code can use a:

IoT Hub lets any device connect without any security:

IoT Hub names must be unique:

You can use Serverless code to respond to IoT events

When you send IoT events to IoT Hub:

When reading events off IoT hub, you can only do so using code that runs in the cloud

Azure Functions can be run and debugged locally:

Serverless code can only be written in JavaScript and COBOL:

When deploying a Functions App to the cloud, you need to create and deploy:

IoT devices are always secure

There are no documented cases of hackers using an IoT device to hack into a network:

You can share your IoT device connection string with anyone

Symmetric key encryption compares to asymmetric key encryption in which ways:

Self-signed X.509 certificates are ideal for a production environment

X.509 certificates:

Your location can be defined using

The types of sensors that can track your location are called:

There is no value in tracking vehicle locations

GPS data is sent from sensors using:

To get a good GPS fix you need to receive signals from at least how many satellites

GPS sensors send data over:

IoT data is stored by IoT Hub

The service to draw maps on a web page is called:

Azure maps plots data using:

Blobs can be retrieved via a URL

GPS coordinates can be used to check if something is in a defined area

GPS is incredibly accurate so can indicate with precision of less than 1M when a device enters a given area

Geofences are useful when tracking vehicles to:

To have multiple services pull data from an IoT Hub, you need to create multiple:

The default search buffer for a geofence call is:

Points inside the geofence have a distance:

Cameras can be used as IoT sensors

Fruit can be sorted using cameras

Images-based AI models are incredibly complex and time consuming to train, requiring hundreds of thousands of images:

The technique custom vision uses to train a model with only a few images is called:

Image classifiers can be trained using:

The hardware that allows ML models to be trained quickly, as well as making the graphics on our Xbox look amazing are called:

IoT devices are not powerful enough to use cameras:

Camera sensors use film to capture images

Camera sensors send which type of data

A published version of a custom vision model is called an:

When images are sent for classification, they then become available to retrain the model:

You don't need to use images captured from an IoT device to train the model as the cameras are as good quality as phone cameras:

Edge computing can be more secure than cloud computing.

Running an ML model on an edge device is less accurate than running an ML model in the cloud.

Edge devices always need an Internet connection.

What kind of format or domain do we need for Custom Vision ML models to properly run on an edge device?

What is a container?

How do you do Custom Vision model retraining for ML models deployed on edge devices?

Which part of your IoT application gathers data?

The only outputs of an IoT application are actuators.

Things don't need to connect directly to IoT Hub, they can use edge devices as gateways.

The three components of architecting an IoT application are

The component that communicates between the things and the components that create insights is:

How do time of flight proximity sensors work?

AI models cannot be used to count objects?

IoT and AI can be used in retail for:

Object detection involves:

Object detectors only return one result no matter how many objects are detected

What is the best domain to use in Custom Vision for stock counting

At least how many images do you need to train an object detector?

IoT devices are not powerful enough to use object detectors

Object detectors give you:

Object detectors can be used to detect where missing stock should be to allow robots to automatically stock shelves

To count stock you only need to consider the count of objects detected by the object detector

Bounding boxes use:

Can detected objects overlap?

IoT devices can be used to recognize speech:

Voice assistants should send all the audio they hear to the cloud for processing:

To recognize speech, IoT devices need large microphones:

Microphones are what type of sensor?

Sound waves are converted to digital signals using:

1 second of 16-bit audio sampled at 16KHz is how large?

Language understanding involves looking for fixed words:

Language understanding involves:

Cloud providers have AI services that can understand language:

Sentences are understood by being broken down into:

The Microsoft service for language understanding is called:

In the sentence 'set a 3 minute timer' the:

Speech generated by AI models sounds monotonous and robotic

AI models can only create speech in American English:

AI models would convert 1234 into which spoken phrase: