Big Data, Data Analytics, Machine Learning, Artificial Intelligence (AI), Data-driven, Data Mining, Deep Learning, Data Science… These are only a few examples out of thousands of concepts and methodologies supporting a powerful trend present nowadays in almost every aspect of our daily lives.
Everyone uses these concepts either in casual or professional conversations and not always in an accurate and proper way. Hopefully, this article will shed light on the most commonly used ones in a brief and easy way no matter the level of expertise:
Data Science can be considered the wider concept in this regard, encompassing all the other related disciplines and concepts. It is based on the application of maths and statistics in a domain of expertise, combined with a problem-solving approach to extract information and insights.
Big Data is a concept that not only refers to a large volume of data to cope with, being this only one of its features. In addition, it has to be considered as well the wide variety of sources of data available in the analysed domain: they can be structured (databases) or unstructured (pictures, video, audio, social media, etc.). The last, but not least, characteristic of Big Data is the high updating velocity of the data, and in consequence, the high velocity needed to process them.
Data Science can be outlined as a process, being the early stages gathering, classifying, cleansing, and organising the data and sources of data available. One of the next stages would be to apply Data Analysis whose objective is to find answers, solutions, and reach conclusions in order to face the existing issues.
Data Analysis is nothing new and it has been performed for decades in any kind of business: using descriptive statistics, data and drawing conclusions. The breakthroughs are indeed the new arising methodologies or applying Data Analysis to Big Data sets (Big Data Analysis).
Data Mining is one of this new techniques. It consists in identifying hidden patterns and relationships within data to predict future trends in the domain of study. It sounds like ‘looking for a needle in a haystack’ and actually is, as there is an intensive workload behind. If only it can be performed in an automated way…
Machine Learning uses algorithms on the data to learn from them, and then predict future trends for a specific topic. These algorithms comprise statistical and predictive analysis to spot hidden patterns and insights based on perceived data. So yes, this is more than performing Data Mining in an automated way.
All of the above have plenty of possibilities and areas of application. However, one of the most interesting fields in this regard is Artificial Intelligence. Artificial Intelligence (AI) is a discipline oriented to develop machines able to carry out tasks that are characteristic of human intelligence. Those include a vast number of functions like language understanding, sound and images recognition, learning, problem-solving, etc.
Narrow AI is focused on a specific capacity of human intelligence whereas General AI includes all the capacities of human intelligence. Due to that Narrow AI is the category with the most presence nowadays. For instance, a machine able to recognize voice commands but nothing else would be an example of Narrow AI.
According to its definition, AI can be developed without using Machine Learning, but it would require an intensive labour of coding complex flow charts based on rules and constraints which have to be analysed in advance.
Nevertheless, applying Machine Learning to build AI is a far better approach. It allows training the algorithms by feeding them with huge amounts of data so they adjust themselves and improve. It can be said that algorithms learn for themselves rather than being taught to perform specific tasks. This speeds up and makes easier the process.
Deep Learning can be considered as an advanced approach to Machine Learning. It is inspired on the structure and way of working of the human brain, based on the interconnection of lots of neurons. Artificial Neural Networks are structured as a set of multiple layers, running in each of them a set of Machine Learning algorithms oriented to a specific human intelligence capability.
‘Data Mining is an advanced way of performing Data Analysis.’
‘Big Data is the resource and Data Mining is the driver to get benefits from it.’
‘Machine Learning is a way of building Artificial Intelligence (AI).’
‘Deep Learning is a particular approach to Machine Learning.’
How are these fields of knowledge applied in Aerospace? Hopefully, the following articles will give you more than a proper answer to that question: