Machine Learning As a Service — How to Get Started

Lili Török
Small Business, Big World
4 min readAug 10, 2018

Heard all about machine learning, but don’t know where to start? Luckily, you don’t have to be a tech genius to deploy machine learning applications at your business.

You may not even have to hire one at first. Several companies offer Machine Learning-as-a-Service apps for businesses that want to try the technology but don’t know where to start.

What Is Machine Learning

According to Stanford University, “machine learning is the science of getting computers to act without being explicitly programmed.”

Computers have always been able to collect and analyze large piles of data within a very short time. However, with machine learning, they’ll now be able to draw conclusions and modify their response accordingly. Just like humans.

There are two main types of machine learning: supervised and unsupervised. Supervised machine learning doesn’t mean it needs human direction and control; it simply indicates that the learning happens through previously structured and labelled data.

For example, if you have a bunch of emails labelled “spam,” and others labelled “important,” your supervised machine learning app can categorize incoming emails according to those labels it learned.

Unsupervised learning means that your app learns from a pile of unstructured and unlabelled data, finding the hidden pattern and teaching its characteristics to itself.

Machine Learning As a Service

There are five major tech providers offering machine learning apps to their customers. How major are they? You must have heard of at least four of them. In fact, you’re most likely already using their services in other areas.

Google

The tech giant offers Google Cloud Machine Learning Engine to its customers. It provides a flexible service, allowing developers to create models for data packages of any size.

The service is integrated with all other Google products. You can try it for free, and you even get a $300 credit. After that, it offers personalized pricing on a pay-as-you-go basis.

Amazon Web Services

Amazon’s machine learning platform is great for beginners. You can use data from many sources, including all other Amazon services, and build your model app according to your needs.

Pricing follows a pay-as-you-go model. Amazon charges $0.42 per hour for data analysis and model building, as well as $0.10 per 1,000 batch predictions, and $0.0001 per each real-time prediction.

Microsoft

If Google and Amazon are there, would Microsoft dare miss out? Nope. Microsoft offers Azure Machine Learning Studio to its clients. The service is very scalable, fitting the needs of tech-savvy customers as well as complete beginners in the field.

Azure is browser-based, allowing you to simply drag-and-drop elements you need into your app. You can realize your machine learning idea within minutes, without prior coding experience.

The basic version of the service is free, but for $9.90 a month, you gain access to the enterprise-grade workspace with lots more support.

IBM

Who’s missing? Right, IBM. The tech giant’s Watson Machine Learning service allows developers to deploy and operationalize machine learning apps. IBM offers visual modelling tools to aid customers in recognizing patterns.

You can try WML for free for 30 days. After that, you can choose between three packages: Lite (still free), Standard ($0.500/1,000 predictions), and Professional ($1,000/instance).

Plus, it beat Ken Jennings in Jeopardy! Impressive.

BigML

Finally a company you’ve (most likely) never heard of. Although BigML doesn’t come in the umbrella of a well-known tech enterprise, it still offers you the same convenience, if not more. You can import data from all sources (including your accounts at the previously mentioned four companies), and enjoy flexible deployment options for any app size.

Make use of BigML’s free models and datasets, as well as visualization tools and clustering algorithms.

Pricing depends on the type of your subscription. You can enjoy a free personal account, or get started with a $30/month standard organizational account and add on extra services as you go.

And since we’re talking about pricing, let us mention a less-known fact. Even though you may not think of them as such, global payments are a service too. Why not choose the best provider in the field?

With Veem, you can send and request global payments with just a click. Track all your transactions in real-time and save money on $0 wire fees, as well as competitive foreign exchange rates.

Sign up for a Veem account to enjoy top notch global payment services.

This content was originally published on the Veem blog. For more information and exclusive content, click here

--

--