Machine Learning Tools Destroy The Passive Process

20th June, 2017 by

Machine learning tools have the potential to change our lives. Many people might argue they already have. When a computer can learn from past experience and adapt its behavior accordingly, it ceases to be a passive processor of instructions. Machine learning involves a degree of judgment that helps computers to make decisions, process data or predict patterns of behavior.

Although it might not be instantly evident, machine learning is all around us. This evolutionary process underpins many of the digital services we’ve come to rely on from spam filters and voice-controlled personal assistants to the curated suggestions provided by streaming services and online retailers. The arrival of high-speed internet access and cloud computing has enabled computer programs to assimilate huge volumes of data, in tandem with ever-more powerful CPUs and GPUs.

The IoT meets Machine Learning

The Internet of Things will further accelerate the twenty-year development of machine learning. Vast quantities of relatively mundane information are uploaded into databases that require processing and analysis to be of any value. Here in 2017, there are already numerous ways machine learning tools can be deployed to the greater good. We’ve listed eight of the most valuable or significant examples below, as well as briefly discussing their optimal uses…


Arguably the best-known machine learning platform, Google’s Brain Team launched TensorFlow as open source just 18 months ago. Having been developed in-house to detect and determine patterns and correlations, it’s compatible with platforms as diverse as 64-bit Linux, iOS and Windows.


Developed in a New Zealand university, Weka features Java algorithms designed to specialize in data mining. Unlike the other machine learning tools in this list, its developers have published a book of practical tips and techniques that can be used to understand its potential. This makes Weka a great place for beginners to start.  


Community-based and non-commercial, Shogun’s origins date back to the last century when it was created for bioinformatics. It’s since been used extensively for scientific research, and is part of NumFOCUS – the open-source computing non-profit used to undertake reproducible scientific research.

Google GoLearn

It’s ironic that a search for GoLearn in parent company Google’s search engine displays few results. Despite this, GoLearn is one of the more high-profile machine learning tools in existence. Reminiscent of Python’s scikit-learn, it uses Google’s Go language to undertake highly customizable data handling.

Amazon Machine Learning

Buoyed by the huge success of their Web Services division, Amazon’s latest venture is aimed squarely at developers. Using visualization tools and wizards instead of algorithms, AML can run regression, classification and categorization tasks. It’s an obvious partner to Amazon Redshift PaaS and MySQL.

Accord Framework

Developed for .net and written in C#, Accord is very useful for signal processing. This is the technology behind facial recognition – rumored to be the future of payments, given the rapid expansion of biometric ID technology. Accord is suitable for clustering, hypothesis testing, classification and regression.


Apache Spark’s scalable machine learning library has been designed to plug into Hadoop workflows for high-speed iterative computations. The nature of Spark means suggested improvements to its algorithms are welcomed, while the MLbase project can sit on top of MLib to answer queries through a declarative language.


The tool to call on for predicting trends or detecting fraudulent behavior, H20 is compatible with everything from Python to Scala. Since it was written in Java, H20 is particularly well suited to Hadoop and Spark. A straightforward GUI simplifies usage; stellar client bases include: eBay, PayPal, Macy’s and Capital One.

The opportunities are endless and the opportunities exciting, but the need for care, security and attention to detail has never been more important. Data is valuable and criminals want it. If you are harvesting and using data imagine you are handling DNA. You have a duty of care to protect the information we all give often with little thought. Security, morality and protection are key watchwords as we develop our lives in the machine learning age.

(Visited 15 times, 1 visits today)