• #AI
• #AIWITHOPENTEXT

AI

Put simply, AI involves teaching computers to do what comes naturally to humans.Humans bank experiences and use this knowledge and the lessons learned to help make decisions. Where humans can instinctively see patterns in relatively small amounts of data, AI has the potential to examine almost any amount of data or variety of data quickly and efficiently to extrapolate meaning and predict outcomes.

While AI has, over the years, conjured the image of super-computers that take over the world, this remains firmly in the province of science fiction. AI does not yet have cognitive capabilities similar to humans, such as making complex judgments, having emotional responses, or handling ambiguity. Businesses today are using machine learning methods to teach their AI how to complete specific tasks without being explicitly programmed. This type of AI allows a user to train the machine to help find the best solutions and focus on higher-value business issues.

Finally, it’s important to remember that AI is not a single technology. It’s a range of technologies that bring capabilities to problem-solving, decision support and task automation.

AI, Now & the Future

When asked, global CIOs placed AI at the top of the list of game-changing technologies, with analytics a close second. These two interrelated technologies are ideally suited to address a significant challenge facing the enterprise: Data.

Data has become a business’s most valuable asset, more useful, indeed, than oil.

However, that value can only be harnessed if an organization uses all the data available. The amount of data is growing exponentially. There is also an increase in the variety of data, which can be broken down into two categories:

Structured data.

Any data resides in a fixed field within a record or file, such as a relational database or spreadsheet.

Unstructured data.

Any data that doesn’t have a pre-defined data model is not organised pre-defined, such as text, video, email, and social media posts.

The speed at which data is created is accelerating rapidly. Every organization is now faced with a virtual tsunami of data. This chapter discusses how AI and analytics to help provide control over business data.

Machine learning

Machine learning is an aspect of AI that uses statistical techniques to give computer systems the ability to learn from data without being programmed to do so. It uses algorithms to parse data, learn from it and then make recommendations or predictions about an action or outcome. In this context, we are learning means the ability to improve or refine performance and accuracy due to the system's data, either structured or unstructured data depending on the system’s level of complexity. Computers require a lot of data to learn, so big data and machine learning become the ideal bedfellows. It can take some machine learning applications a vast amount of separate data points and datasets to accurately and adequately learn the task it is to accomplish.

In practice, machine learning software consists of algorithms that can learn from and make predictions based on the data it receives. It can go beyond simply following static program instructions by building models from sample inputs to help the system make data-driven predictions to support decisions and actions.

A brief word on data preparation and integration

In some ways, the most challenging task within data analytics is not the analysis itself. It begins with data collection and preparation. Users must be sure they have all the data required, and it is up-to-date and accurate and in a useable form for systems.

Users need to identify the business issue they seek insight into and the required data to reveal that insight within data collection. Data from different sources must be combined via data integration and transformed into a standard format that can be loaded into the analytics system.

Once the data required is in place, the next step is to fix data quality problems that could affect the accuracy of analytics applications. This will include data harmonization, data profiling and data cleansing to ensure that the information in a dataset is consistent and that errors are eliminated. In addition, data governance policies need to be applied to ensure the data is adequately used and complies with corporate policies and data regulations, such as the EU’s General Data Protection Regulation (GDPR).

Only after this point is reached can the real work of AI and analytics begin in earnest.