AWS ML Services

AWS Machine Learning tools provide a number of high-level algorithms that provide business intelligence across various data sources including text, images, and video. We are at the beginning of what will become a machine-learning revolution, where businesses that have embraced machine-learning early will outperform their competition.

"Machine Learning will empower and improve every business, every government organization, every philanthropy - basically there's no institution in the world that cannot be improved with Machine Learning"

bezos
FOUNDER, AMAZON Jeff Bezos

When it comes to artificial intelligence, the buzzword du jour is Generative AI, but Machine Learning is the technology that underpins all of it (seen in the diagram to the right). GenAI provides a selection of models that be tasked with activities, and Machine Learning is the system that can be used to create new models. We aren't suggesting creating new models—there are already over 1.6 million on Huggingface—but some tasks like product recommendations are better suited to and more cost effective to use machine learning for.

With that in mind, there  total of 33 AWS machine learning services, not including their latest GenAI services. Within this broad umbrella of machine learning, there are five major categories:

  1. AWS Supported ML Frameworks
  2. AWS Deep Learning Algorithms
  3. Add on ML Services
  4. ML Powered Hardware
  5. Special AWS ML Services
artificial intelligence diagram-color2

1. AWS Supported Machine Learning Frameworks

Among the AWS machine learning services offered, the machine learning frameworks are the most rudimentary. AWS provides the hardware and optimizes performance for the following write-your-own-algorithm frameworks:

  1. Amazon SageMaker - Amazon's very own machine learning framework.
  2. PyTorch on AWS -  a machine learning framework managed by Facebook's AI Research (FAIR) Lab.
  3. Apache MXNet on AWS - a machine learning framework from the Apache Software Foundation.
  4. TensorFlow on AWS - a machine learning framework managed by the Google Brain team.

In addition to these AWS also does support the following via their available deep-learning Amazon Machine Images (AMIs):

  • Chainer
  • Theano
  • Keras
  • Gluon

2. AWS Deep Learning Algorithms

The most robust offering —and by far the most interesting— is the AWS deep learning algorithm which spans a large cross-section of data and brings a tremendous amount of value with no need for any kind of training. These deep learning algorithms include:

  1. Amazon Comprehend - discover insights and relationships in text
  2. Amazon Comprehend Medical - a medical-specific spinoff of Comprehend
  3. Amazon DevOps Guru - ML-powered cloud operations service to improve application availability
  4. Amazon Forecast - increase forecast accuracy using machine learning
  5. Amazon Rekognition - machine learning computer vision to analyze image and video
  6. Amazon Personalize - create real-time personalized user experiences faster at scale
  7. Amazon CodeGuru - automate code reviews and optimize application performance with ML-powered recommendations
  8. Amazon Fraud Detector - a real-time fraud detection service
  9. Amazon Kendra - an intelligent search service powered by machine learning
  10. Amazon Textract - extract printed text, handwriting, and data from any document
  11. Amazon Translate - translate written text from one language to another
  12. Amazon Transcribe - convert spoken language into written text
  13. Amazon Lookout for Equipment - detect abnormal behavior by analyzing sensor data
  14. Amazon Lookout for Metrics - detect anomalies in metrics
  15. Amazon Lookout for Vision - spot product defects using computer vision to automate quality inspection
Machine Learning

What is Deep Learning?

Deep learning is a category of AI used to find patterns within various types of complex datasets which can include textual, audio, and visual data.

3. AWS Machine Learning Add-on Services

Not exactly stand-alone products (but branded that way), the machine learning add-ons category includes offerings that generally make some of the drudgery involved in machine learning less painful or somehow improve performance. These services include:

  1. Amazon Augmented AI - easily implement a human review of ML predictions
  2. Amazon Elastic Inference - lower machine learning inference costs by up to 75%
  3. Amazon SageMaker Ground Truth - create datasets for training machine learning models
  4. Amazon SageMaker Neo - run ML models anywhere with up to 25x better performance
  5. AWS Deep Learning AMIs - Amazon Machine Images (AMI) for different ML frameworks
  6. AWS Deep Learning Containers - read-to-go containers for different ML frameworks
  7. Amazon HealthLake - Securely store, transform, query, and analyze health data in minutes
Machine Learning

What is Machine Learning?

Discover the basics of machine learning through a practical analogy, covering data gathering, cleaning, hypothesis formation, training, and...

EBOOK

Get our free ebook on AWS Machine Learning

Our AWS Machine Learning Survival Guide provides the key insights you'll need to successfully navigate the adventure of ML implementation.

 

GET THE EBOOK
AWS-Machine-Learning-Survival-eBook-(1)-1

4. AWS Machine Learning Powered Hardware

  1. Amazon Inferentia - high-performance machine learning inference chip, custom designed by AWS
  2. AWS DeepLens - a deep learning-enabled video camera
  3. Amazon Monitron - an end-to-end system for equipment monitoring (including a physical sensor)
  4. AWS Panorama - hardware-enabled computer vision at the edge

Special AWS Machine Learning Services

Not exactly deep learning algorithms, there are some AWS machine learning algorithms that really are their own strange little thing.  These include:

  1. Amazon Lex - build voice and text chatbots
  2. Amazon Polly - turn text into life-like speech
  3. AWS DeepRacer - autonomous 1/18th scale race car, driven by ML

Why use AWS ML Tools?

While it is possible to build your own algorithm using open-source deep learning frameworks like MXNet, Keras, and TensorFlow most problems or opportunities presented by data are not unique. They are problems that other organizations have seen and solved. Some examples of this are:

  • What products are individual customers most likely to purchase as an upsell?
  • For content being posted by users can we use a machine to tell if it is objectionable (nudity, profanity, etc.)?
  • Is there a way we can pull out themes and sentiments from large amounts of customer product reviews?

The build-your-own solution requires massive amounts of trial and error, overseen by teams of data scientists. Thankfully, this work has already been done by AWS and high-level algorithms (in these cases Personalize, Rekognition, and Comprehend) can simply be implemented and trained at a fraction of the cost and time.

Get Started with AWS ML

We build the most valuable, safest solutions for our customers and have a significant amount of experience across all the major public Cloud providers. 

Schedule a Free Consultation

Speak with our team to understand how Metal Toad can help you drive innovation, growth and success. 

GET IN TOUCH

Learn more about machine learning from our blog

Machine Learning

Is Machine Learning Artificial Intelligence?

Gain clarity on the relationship between Artificial Intelligence (AI) and Machine Learning (ML). Discover the role of AWS in accelerating AI and ML...

Nathan Wilkerson, VP of Engineering Feb 21, 2024