AI and ML

Artificial Intelligence & Machine Learning

Artificial Intelligence (AI)

Wikipedia defines AI as “intelligence demonstrated by machines, in contrast to the natural intelligence displayed by humans”. It is also defined as “to simulate any intellectual task”. These definitions give a very wide scope of what AI means. A few examples are:

• Search and planning (e.g. playing chess)
• Perception
• Reasoning and knowledge representation (e.g. Watson on Jeopardy)
• Ability to move and manipulate objects
• Natural Language Processing (communication)
• Learning

Historically, AI has been associated with programming things that machines can do. Though this kind of pre-programming nowadays doesn’t seem like AI to us. What seems more AI is that the machine can learn by itself “through repeatedly observing what humans do”, that seems like AI.

Machine Learning (ML)

Machine learning is when machines perform a specific task without using explicit instructions, but by relying on patterns and inference instead. This is a subset of AI, but one that is progressively encroaching into the space of other subsets.

For example, for long object recognition has been done by complicated filters using image processing technologies, but in recent years this has completely changed by ML methods of Convolution Neural Networks (CNNs). Similarly, speech recognition, OCR, recommendation systems, certain medical diagnosis, spam filtering, search engines, etc. all these fields of AI have totally transformed by use of ML methods.

Machine learning is all about recognizing patterns, making predictions and by learning based on how well the prediction was from real data. This leads us to our next big chunk : Data Science

Data Science

Once again wikipedia defines Data science as “concept to unify statistics, data analysis, machine learning and their related methods” in order to “understand and analyze actual phenomena” with data. It employs techniques and theories drawn from many fields within the context of mathematics, statistics, computer science, and information science.

The usage of the term data science ranges from data mining (discovering patterns in large datasets), big data (analyzing complex and extremely large datasets), data wrangling (transforming/mapping raw data into another format), statistics (collection, organization, analysis, interpretation, presentation of data), probability (numerical description of likelihood or certainty), etc.

Here again, if you are alert you can see the huge overlap of machine learning with data science.