Machine Learning

What is Machine Learning?

The crux of Machine Learning (ML) is based on the recognition that human intelligence hinges more on ‘probability’ than ‘reason’ or ‘logic’. Our brains unconsciously perform more estimation – calculating the fastest driving route based on prior experience, guessing an opponent’s move while playing a board game, plotting a strategy – than using the skill of reasoning and thinking. The human brain relies first and foremost on its ability to assess how likely something is; we are not consciously aware of this.

The machine is flexible in its learning and uses an iterative approach to independently adapt to new data. The models learn from previous computations to produce repeatable and reliable results and decisions. Machine Learning algorithms have been around since the 1950s. However, the ability to apply complex mathematical calculations to big data is a recent development. The growing volumes and varieties of available information, cheap and powerful computational processing and affordable data storage makes the comprehensive machine learning capabilities available.

Examples of Machine Learning applications:

Self-driving cars Speech recognition Effective web search Fraud detection.
Machine learning is so common that we use it dozens of times a day without even knowing it.

All of these things mean it's easier than ever for organizations to benefit from this technology by identifying profitable opportunities or avoiding unknown risks, building precise models and analyzing them to deliver effective solutions.

Machine Learning Use Cases:

Amazon and Netflix:
  • Perhaps the most famous use cases of machine learning are the algorithms associated with Amazon and Netflix. Both use searches, customer feedback and previous customer history to provide a personalized experience and recommend content and products the customer may enjoy.

  • Financial Industry:
  • Fraud detection and prevention is one of the major benefits of machine learning in the financial industry, but one other aspect is its ability to minimize false-positives. Just as consumers are thankful when fraud is prevented, they are equally annoyed when their credit cards are flagged for fraud when it shouldn’t be. Machine learning can update its algorithms to better indicate when a purchase is fraudulent versus legitimate.

  • Healthcare:
  • Machine learning algorithms have proven to spot patterns in x-rays and MRIs that humans could not. In preventative care, such as routine mammography, machine learning algorithms identified over half of the cancers nearly a year before an official diagnosis.

  • Machine learning has also been used by government authorities like the Center for Disease Control and Prevention to understand, spot and manage risk factors for diseases.

One of the critical questions about machine learning is, how much of human intelligence can be approximated with statistics?