You work for an auto parts manufacturer as a machine learning specialist.
You need to build a machine learning model that categorizes proprietary auto parts as they traverse your plant's production lines.
You do not have any existing trained models from which to start your work.
You plan to use an image classification algorithm such as ResNet to classify the auto parts with one or more labels.
The classified image data will then be used by your accounting department to dynamically keep the company's parts database updated with the newly produced units. Since you are building a model to classify images of proprietary auto parts, which technique can you use within SageMaker to expedite the deployment and operation of your model?
Click on the arrows to vote for the correct answer
A. B. C. D.Answer: C.
Option A is incorrect.
Online learning refers to the process of training your model incrementally by giving it data observations as individual observations or in mini-batches.
This will train your model.
But it won't expedite the process.
Option B is incorrect.
Incremental learning would help expedite the training process if you start with an existing model and extend it with new data, specifically your proprietary auto parts images.
However, you don't have any existing trained models from which to start your work.
Option C is correct.When you use transfer learning, you start with an ‘off the shelf' trained model from a source such as ONNX Model Zoo.
You take the off the shelf trained model and apply it to your different but closely aligned observations.
This saves you time deploying and operationalizing your machine learning solution since you start from a pre-trained model.
Option D is incorrect.
Out-of-core learning is used to train huge datasets that you can't load into your server's memory.
This algorithm loads some of the data, trains on that subset, loads another subset of observations, trains on that subset, and repeats this process until it has completed the training of all the observations.
This process will not help you deploy and operationalize your model more expeditiously.
Reference:
Please see the Amazon SageMaker developer guide titled Using Amazon SageMaker Built-in Algorithms, the Amazon SageMaker machine learning blog titled Now easily perform incremental learning on Amazon SageMaker, and the article titled Transfer learning with MXNet Gluon.
The technique that can be used within Amazon SageMaker to expedite the deployment and operation of the machine learning model for image classification of proprietary auto parts is Transfer Learning.
Transfer learning is a technique used in machine learning where a pre-trained model, such as ResNet, is used as a starting point to train a new model for a different but related task. In this case, the pre-trained ResNet model can be used as a starting point to classify the auto parts on the production line, with one or more labels.
By using transfer learning, the model can be trained much faster than starting from scratch since the pre-trained ResNet model has already learned to recognize common features and patterns in images. The model can then be fine-tuned with the proprietary auto parts images to recognize the specific features unique to them.
Amazon SageMaker provides built-in algorithms that support transfer learning, such as the image classification algorithm. The pre-trained ResNet model can be easily loaded into the algorithm and fine-tuned with the proprietary auto parts images.
Using transfer learning not only expedites the deployment and operation of the model, but also improves the accuracy of the model since it leverages the knowledge learned from a large dataset of similar images.
Therefore, the answer to the question is C. Transfer learning.