Select Page

Last week we offered a simple definition of AI as the science of making machines complete tasks that require intelligence when done by humans. These days, when people discuss AI, you will often hear the term Machine Learning. What exactly is Machine Learning? Why is it useful? Is Machine Learning synonymous with AI? And if not, how are the concepts unique? First, let’s define Machine Learning.

What Really is ML?

Machine Learning is an AI technique pertaining to computer systems learning on their own, i.e. using data to infer patterns to predict outcomes. The term was coined by computer scientist Arthur Samuel in 1959. Machine Learning is a great method for data analysis and helps to automate analytical model building. A good example would be spam email: Machine Learning analyzes past spam to better identify future junk email.  Data scientists use Machine Learning to fill in existing gaps in data sets. In this way, machines emulate the human brain. When we solve riddles, for example, we look for hidden clues first, then aggregate all the pieces of information together. Our brains draw inferences across this data and using these connections, we can complete the existing puzzle.

Left to Right: 1. Arthur Samuel plays checkers with an IBM 704 computer in Poughkeepsie, New York. 2. The term Big Data, probably originated in the lunch-table conversations at Silicon Graphics in the mid-1990s, in which John Mashey figured prominently.

Why is ML Useful?

Why would you need a process like Machine Learning? To parse through enormous volumes of data. Another term that has taken up residency in 21st century parlance is Big Data. Big Data is defined by the four V’s: variety, velocity, veracity, and volume of data that would be impractical for a human to sift through on his own. Particularly within today’s synergistic, corporate world, Machine Learning and Big Data are often cited by companies looking to demonstrate they operate on the vanguard of technology. Organizations will use AI – particularly Machine Learning – to train a machine with a set of instructions and allow the machine to run on its own, completing the process thousands or even millions of times across massive sets of Big Data, generating results within seconds. For instance, financial services companies use Machine Learning to sort through Big Data and predict future trade prices of stocks and bonds.

The Four V’s of Big Data. Source: IBM

How is AI Unique from ML?

It is best to think of AI as tools and goals. Goals are results driven and tools are the instruments to get you there. Machine Learning is only one tool within AI; AI is a field encompassing many different tools, with each tool targeted towards producing a specific goal in a specific domain, e.g. Natural Language Processing would be a suitable AI tool to achieve an AI goal of sentiment analysis. AI goals can include everything from predicting consumer buying behavior to image recognition to determining whether network traffic is benign or malicious. We use eminent AI tools to achieve these goals. Some of these AI tools include:

  • Machine Learning – Classification
  • Machine Learning – Clustering
  • Natural Language Processing
  • Decision Theory
  • Game Theory
  • Search & Optimization
  • Markov Decision processes
  • Deep Learning

Some of Avata’s AI Tools

Avata Intelligence’s AI solution is a consummate AI solution: rather than implementing one or two of these sophisticated AI tools, the Avata platform implements the full scope of AI facility to best analyze data and make recommendations.

Going Forward

As we move forward as both a culture and a civilization, it is fundamentally important we utilize computers and AI mutualistically to make the best use of our resources. Humans and machines working in tandem provides the best potential quality of life. One such usage is self-driving technology in cars. Sensors along the exterior of the vehicle offer 360° of awareness. This environmental data is channeled through the automotive AI, and determinations of speed, acceleration, acceptable times to pass, etc., are made faster and with less rate-of-error than cars with human drivers. (Kang, Cecilia. (2016, September 19). Self-Driving Cars Gain Powerful Ally: The Government. The New York Times. Retrieved from http://www.nytimes.com)

Next week we’ll discuss how the combination of these AI tools is utilized to overcome complex challenges.