Last updated on August 6, 2018.
Published on August 6, 2018.
Forward-thinking organizations are increasingly including Machine Learning capabilities as part of their overall digital business strategy. When used effectively, Machine Learning can enhance existing products and services with new smart features, or reduce operating and production costs.
As organizations proceed from strategy development to implementation, Machine Learning platforms and their services support numerous intelligent applications for lightweight, agile POCs as well as highly scalable production deployments, with all the benefits expected from a cloud-based infrastructure.
The top cloud platform providers we evaluated are Amazon, Google, IBM, Microsoft, Oracle, Salesforce, and SAP. We assigned the intelligent services from each cloud provider to a service category when the services offered similar capabilities. Please see the table below.
Machine Learning Cloud Platform Comparison
| Service | AWS | IBM | Microsoft Azure | Oracle | Salesforce | SAP | |
|---|---|---|---|---|---|---|---|
| Machine Learning | Amazon Machine Learning | Cloud Machine Learning Engine | Watson Machine Learning | Machine Learning Studio and Machine Learning Services | Oracle AI Platform Cloud Service (Coming Soon) | Salesforce Einstein | SAP Leonardo Machine Learning Foundation |
| ML Model Builder | Amazon SageMaker | ||||||
| Suite | Cloud AI, Cloud AutoML (Beta) | Watson Studio | SAP Leonardo Machine Learning Foundation | ||||
| Image / Video Analysis | Amazon Rekognition, AWS DeepLens | Cloud Vision, Cloud Video Intelligence | Watson Visual Recognition | Computer Vision API, Video Indexer, Custom Vision Service (Preview) | Einstein Vision | SAP Leonardo Machine Learning Foundation | |
| Image / Video Search | Bing Video Search API, Bing Image Search API | Einstein Vision | SAP Leonardo Machine Learning Foundation | ||||
| Personality / Emotion Detector | Watson Personality Insights | Emotion API (Preview) | |||||
| Face Detector | Amazon Rekognition, AWS DeepLens | Cloud Vision | Face API | SAP Leonardo Machine Learning Foundation | |||
| Pre-config. Environment | Amazon Deep Learning AMIs | ||||||
| Training Framework | Apache MXNet on AWS | ||||||
| Open Source Deep Learning | TensorFlow on AWS | ||||||
| Content Analysis | Watson Discovery | Content Moderator | SAP Leonardo Machine Learning Foundation | ||||
| FAQ Builder | QnA Maker API | ||||||
| Catalog | Watson Knowledge Catalog | ||||||
| Job Discovery | Cloud Talent Solution |
All providers offer a scalable, high-performance Machine Learning service for creating models and generating predictions. Amazon offers a product, Amazon SageMaker, in addition, which is specifically geared toward building Machine Learning models. With Amazon SageMaker you can choose any framework or algorithm, train and deploy models, and integrate with existing workflows.
Google, IBM, and SAP offer a suite of products. A suite is a combination of tools and services bundled together and sold as one. Google’s Cloud AI suite includes Cloud AutoML, Cloud TPU, Cloud ML Engine, and more. IBM’s Watson Studio provides a suite of tools to collaborate and work with data to build, train and deploy models. SAP Leonardo Machine Learning Foundation includes services for image processing, natural language processing, time-series processing, and custom ML models.
Many companies also offer services for image and video analysis for classifying images, detecting objects and faces, gaining insights, moderating content, and extracting text. Some offer personality/emotion and face detectors, which are beneficial in finding psychological traits, personalizing the user experience, and organizing and tagging faces in photos. IBM and Microsoft also offer specific services for moderating text and images, content analysis, through Machine Learning.
Unique offerings:
- Amazon offers a training framework, Apache MXNet on AWS, an open-source Deep Learning framework, TensorFlow on AWS, and pre-configured environments for deep learning, Amazon Deep Learning AMIs.
- Google offers a service, Cloud Talent Solution, which powers job sites with Machine Learning.
- Microsoft Azure offers an FAQ builder that extracts questions and answers and uses testing and training techniques to publish frequently asked questions.
- IBM offers a catalog, Watson Knowledge Catalog, which is used for accessing, creating, organizing, and sharing data and models.
Click here to contact us if you would like to discuss how these Machine Learning services can be effectively put to use to meet your project needs.