Artificial intelligence (AI) has become a decisive factor for companies to remain competitive now and in the future. Currently, more than half of all German companies are actively involved in AI and especially in machine learning. It is becoming steadily easier for users to get to grips with this topic, and meanwhile smaller companies can also benefit from it by relying on pre-structured "as-a-Service" solutions (MLaaS). We provide an overview of the services offered by the three major cloud providers – Amazon, Microsoft and Google – and how they are being used.
What is Machine Learning as a Service (MLaaS)?
Small and medium-sized enterprises usually lack the resources to develop and implement their own machine learning applications. MLaaS describes the offer to run corresponding applications directly on the platform of an external provider. Based on the principle of Software as a Service (SaaS), it supplies development tools, for example for data visualization or predictive analytics, to develop its own intelligent products and solutions in a time- and cost-saving manner. It provides the data center capacities and takes over infrastructure topics such as data pre-processing, model training as well as the actual data calculation and evaluation with further forecasts. Moreover, the provider is responsible for the installation and configuration of the API, which links the application with the company's own IT infrastructure.
AWS, Microsoft Azure or Google Cloud: partners for implementation
MLaaS offers are available in different versions for the most diverse areas. The majority is provided by the three big public cloud providers Amazon Web Services (AWS), Microsoft Azure and Google Cloud Platform. They cover a wide range of applications and currently offer the best possible service.
Below you will find all relevant MLaaS offerings from AWS, Azure and Google and their specific application areas. More details about the individual solutions can be found by clicking on the respective link.
What MLaaS solutions are there?
Software platforms for artificial intelligence (AI) enable users to develop and train machine and deep learning models in a time- and cost-saving manner. They provide pre-developed models that can be used via drag-and-drop and individually optimized or trained for the required tasks.
Natural Language Processing (NLP)
Due to the rapid development of deep learning NLP has made enormous progress in recent years and enables applications to interact with human language. Its operational areas cover among other things the machine translation, grammatical syntax analyses, tendency analyses and part-of-speech-tagging as well as the assignment of words and punctuation marks of a text to certain word kinds. It is therefore the most important technology behind chatbots, for example.
- Amazon Polly
- Amazon Lex
- Amazon Translate
- Amazon Comprehend: NLP
- Azure Language Understanding (LUIS)
- Azure Bing Spell Check API
- Azure Cognitive Services
- Azure Text Analytics API
- Azure Linguistic Analysis API
- Azure Translator Text API
- Google Cloud Translation API
- Google Cloud Natural Language API
This technology enables applications to understand spoken language and convert it into text. Typically, developers combine natural language comprehension and intentional analysis to determine what a user is looking for. Prominent examples include language assistants like Siri, Google Home and Alexa. Speech recognition can also be used for many other purposes, including smart home, therapeutic, language learning, and real-time writing for deaf people.
- Amazon Transcribe
- Azure Speaker Recognition API
- Azure Translator Speech API
- Azure Custom Speech
- Google Cloud Speech-to-Text API
- Google Dialogflow
This deep learning technology enables applications to recognize and understand images and videos within seconds. Computer vision algorithms can analyze visual content to identify and classify specific objects, people, and streaming videos. Companies like YouTube, for example, use it to ensure that no inappropriate content is uploaded. Other applications include signature recognition, visual material verification, and medical image recognition.
- Amazon Rekognition
- Azure Computer Vision API
- Azure Content Moderator
- Azure Custom Vision Service
- Azure Face API
- Azure Emotion API
- Azure Video Indexer
- Google Cloud Vision AI
- Google Cloud Video AI
What is the right MLaaS solution?
Choosing the right solution depends on the requirements and needs of your business. First you should determine in which area an MLaaS solution can support you, and then compare the corresponding offers from the different providers. There's nothing wrong with using multiple applications from different vendors in a multi-cloud strategy recommended by Gartner.