You work as a machine learning specialist for a home maintenance automation company that produces robots to vacuum the floor, mow the lawn, and other automated worker tools.
You have built and trained your model (starting from a pre-trained model from ImageNet) using the SageMaker built-in Object Detection algorithm.
The robots use the Object Detection algorithm to detect objects that are obstacles or boundaries in their work area.
You now need to have the robots run in real home settings using your model.
You also want your robots to communicate with each other if there is more than one robot in the operating area. Which set of Amazon services will give you the most cost-effective solution?
Click on the arrows to vote for the correct answer
A. B. C. D.Answer: A.
Option A is correct.
Amazon Elastic Inference gives you the inference processing (CPU, GPU, etc.) you need to process your obstacle and boundary observations.
AWS IoT Greengrass gives you the capability to run inference on your robot devices and communicate with other IoT devices.
Option B is incorrect.
Amazon Sumerian is used for augmented reality, which is not needed to solve your machine learning scenarios.
Option C is incorrect.
Amazon Rekognition is used for image and video analysis.
It would identify objects in your domain.
But it wouldn't contribute to lowering the cost of your inference implementation.
Option D is incorrect.
Amazon Rekognition is used for image and video analysis.
It would identify objects in your domain.
But it wouldn't contribute to lowering the cost of your inference implementation.
Also, Amazon Sumerian is used for augmented reality, which is not needed to solve your machine learning scenarios.
Reference:
Please see the Amazon SageMaker Overview, particularly the Deploy and manage models in the production section, the Amazon Elastic Inference Overview, the AWS News blog titled Amazon Elastic Inference - GPU-Powered Deep Learning Inference Acceleration, the Amazon SageMaker developer guide titled Object Detection Algorithm, the AWS IoT Greengrass Overview, the Amazon Sumerian Overview, and the Amazon Rekognition Overview.
The best option for a cost-effective solution to deploy the object detection model in the robots and enable them to communicate with each other would be to use AWS RoboMaker and AWS IoT Greengrass. Therefore, the correct option is B.
Here's an explanation for each of the options:
A. Amazon Elastic Inference and AWS IoT Greengrass: Amazon Elastic Inference can be used to optimize the model's performance by attaching GPU-powered acceleration to the EC2 instances that run the inference workload. However, Elastic Inference does not provide any support for communication between robots. AWS IoT Greengrass can be used to enable robots to communicate with each other, but it does not provide any support for object detection.
B. AWS RoboMaker and Amazon Sumerian: AWS RoboMaker is a service that can be used to develop, test, and deploy robot applications. It provides a simulation environment that can be used to test the application before deploying it on the actual robots. RoboMaker supports object detection using the RoboMaker Simulation Engine. Amazon Sumerian is a service that can be used to create and publish AR, VR, and 3D applications. It can be used to create a visual representation of the robot and its surroundings, but it does not provide any support for object detection or robot communication.
C. Amazon Rekognition and AWS IoT Greengrass: Amazon Rekognition can be used to perform object detection, but it is not specifically designed for use in robots. It can be used to detect obstacles or boundaries in the robot's work area, but it may not be optimized for use on the robot's hardware. AWS IoT Greengrass can be used to enable robots to communicate with each other, but it does not provide any support for object detection.
D. Amazon Rekognition and Amazon Sumerian: Amazon Rekognition can be used to perform object detection, but it is not specifically designed for use in robots. Amazon Sumerian can be used to create a visual representation of the robot and its surroundings, but it does not provide any support for object detection or robot communication.
Therefore, the best option for a cost-effective solution to deploy the object detection model in the robots and enable them to communicate with each other would be to use AWS RoboMaker and AWS IoT Greengrass.