Knowledge Mining with Artificial Intelligence

Knowledge Mining with Artificial Intelligence

Artificial Intelligence and Machine learning have gained widespread popularity during the past two decades. Increasingly more companies are seeing the vast number of promising applications that these two technologies have to offer, allowing them to reduce costs, optimize business processes, and increase productivity.

However, up until now the field of Artificial Intelligence and Machine Learning have been focusing on delivering so-called Narrow AI Models. The term ‘narrow models’ is used to denote models which are designed to excel in very specific tasks. An example of such a model is an Artificial Neural Network (ANN) which was specifically trained to detect license-plates from camera footage.

Within business – however – there is a strong urge for machine learning models which are able to carry out tasks with a scope that is much broader than the original task they were designed to carry out. Models which can carry out such a vast amount of (intellectual) tasks are often categorized under the stream of ‘strong’ or ‘general’ artificial intelligence. One of the technologies that is on the frontline of this stream within Artificial Intelligence is Knowledge Mining.

This article will explain what knowledge mining is, how it is different from currently existing techniques, and – perhaps most importantly – what the technology can offer to businesses in general.


Data Mining vs. Knowledge Mining

To explain what knowledge mining is and how we got here, we need to start with the foundations of the technology: Data Mining.

Data mining – a term that was first coined in 1983 by economist Michael Lovell – refers to the practice of using mathematical and statistical models to extract, structure, and transform data so that it can be used for further analysis and modelling. The rationale behind this practice is that machine learning models often require a large number of data points for training – and therefore predictions – being effective.

Data mining is currently being used in a wide range of different industries. Marketing, for example, utilized the technique to gather and analyze relationships between customer parameters and the type of products they buy – allowing them to build customer profiles and engage in highly targeted marketing campaigns.

Another example can be found in the retail industry. Nowadays, supermarkets spend a lot of time, money, and effort into finding associations between different products. The rationale behind this lies in the fact that positioning products which are often bought together in the same aisle may drastically improve the sales of said products.

However, the process of data mining does not provide you with precooked answers on its own. It usually requires a bunch of data scientists and data engineers to build intelligent models on top of the acquired data before the underlying value can be captured by the business.

This is where Knowledge Mining comes in. As the name suggests, instead of mining for a vast amount of data – as is being done with data mining -, knowledge mining primarily focuses on finding intelligent insights in the data that the algorithm is subjected to. Usually, such knowledge is represented in relations or correlations between individual data points and patterns in between different data sources.

In this way, data mining can be considered as the little brother of knowledge mining: whereas data mining takes care of the gathering and the structuring of data, knowledge mining takes care of finding the actual knowledge that lies hidden deep inside the data.


What is in it for you?

So what does this mean for you?

How can knowledge mining be of value in your business?

First of all, implementing knowledge mining techniques will allow you to save a lot of time and resources. A good knowledge mining solution will take over at least part of the task of obtaining business-critical insights. Thereby, the business can effectively reduce the time that analytical-minded people have to scavenge through piles of data, allowing them to put this invaluable human capital to use for other purposes.

In addition, the insights that result from applying knowledge mining techniques within a business environment are invaluable as well. This is shown by a recent study conducted by Harvard Business Review (HBR), which states that around 68% of business leaders consider knowledge mining to be a critical factor in achieving their companies’ strategic goals over the next 18 months (see references). The examples on how this value can be delivered are numerous, but perhaps one of the most straightforward use-cases lies with businesses which have a vast amount of disparate data sources which are currently not being used or are too difficult to search manually.


Challenges in Knowledge Mining

Whereas the business case is obvious in the example above, cautiousness is needed when implementing knowledge mining systems. Just having a knowledge mining engine build on top of your database system might speed up the task of finding information and intelligent insights. However, the implementation of a knowledge mining application should always start with a concrete and tangible business-case for which knowledge mining is the obvious solution.


Conclusion and recap

Machine learning and artificial intelligence have significantly revolutionized the way businesses operate and provide value to their customers. Whereas most AI and machine learning implementations belong to the stream of narrow AI models, there is a significant demand for techniques which go beyond the traditional single-application models and are able to provide a much broader value to the entire organization. Increasingly more machine learning practitioners are looking in the direction of Knowledge Mining for this: a widely applicable technique which has the potential to drastically improve business processes across your companies’ entire value chain.


Data & Artificial Intelligence Team