Can you tell the difference? (AI/ML version)

Can you tell the difference?

AI / ML / Deep Learning

Part 1: [Jargon speak] Defining Smarter Systems

​Artificial Intelligence (AI): Designed to mimic human capacity for learning, perception & interaction

​

​Machine Learning (ML): Designed to 'learn' on its own by observing, i.e., establish correlations & conclude, basis the past, expected outcome or comparable human performance.

​

​Deep Learning: Designed not only to learn on its own from big data, but also improve on existing human/ system performance through neural networks to arrive at more efficient decision rules

Multiple Choice

I wrote 10,000 lines of code. The program can respond & act on things like "What's the weather today?" based on my location or "Add milk to my grocery list". It cannot understand anything new. I have to stick to the verbatim coded script.

Which of the following is the most appropriate when describing my system?

• Artificial Intelligence

• Machine Learning

• Deep Learning

• Its just a manual program

Multiple Choice

I wrote a program that can act as my virtual assistant. I did not hard code every single question & response explicitly.

Which of the following is the most appropriate when describing my system?

• Artificial Intelligence

• Machine Learning

• Deep Learning

• Its just a manual program

Multiple Choice

I wrote a program that can go through my Whatsapp conversations & emulate my response. For example, it responds with a version of "Thank You" & 1-2 of my most used happy smileys to a "Happy Birthday" message. It doesn't understand what it means, just knows what to do basis my past behaviour.

Which of the following is the most appropriate when describing my system?

• Artificial Intelligence

• Machine Learning

• Deep Learning

• Its just a manual program

Multiple Choice

I wrote a program that can understand language. It can read my emails, whatsapp conversations & understand the context to assist me with relevant info on recommended res. I neither programmed explicitly, nor limited where it can learn from.

Which of the following is the most appropriate when describing my system?

• Artificial Intelligence

• Machine Learning

• Deep Learning

• Its just a manual program

How did you do?

​Let's see if we can simplify the English & try again with real world examples

Part 2: [Plain English] What do they really mean?

​Artificial Intelligence (AI): Does tasks typically requiring human intelligence/ Imitates humans

​

​Machine Learning (ML): Learns by experience, i.e., acquires its skills by observing the output when tweaking a set of parameters given to it. Majority of ML is based on statistical modelling, i.e. Pattern Recognition. It can only work with the kind of data it has come across before, cannot deal with unstructured new data

​

​Deep Learning: Inspired by the human brain to learn from large amounts of data. It performs a task repeatedly, each time tweaking it a little to improve the outcome. Used to solve problems that require 'thought', i.e., multiple layers of learning (hence the name neural networks). Can deal with new unstructured data, eventually figures out what to do after a few iterations

Multiple Choice

E-mail service providers like Google & Microsoft use systems that can identify spam without the need for a human intervention. It identifies recurring patterns from decades of spam emails that humans have already flagged and reviewed.

Which smart algorithm/ system is used here? Choose the most appropriate option

• Artificial Intelligence

• Machine Learning

• Deep Learning

• Its just manual

Multiple Choice

IBM's "Deep Blue" can identify chess pieces & knows how each of them move. While the machine can choose the most optimal move from several possibilities, it can’t predict the opponent’s moves.

Which smart algorithm/ system is used by Deep Blue?

Choose the most appropriate option

• Artificial Intelligence

• Machine Learning

• Deep Learning

• Its just manual

Multiple Choice

Which smart system is used by Siri/ Cortana/ Google Assistant? Choose the most appropriate option.

(Hint: Remember that it can understand language, choose the kind of response to our interaction, take a converstation forward, etc.)

• Artificial Intelligence

• Machine Learning

• Deep Learning

• Its just manual

​While these terms may seem abstract at first glance, with the help of Venn diagrams, they can be understood more clearly. In summary, artificial intelligence is the collection of tasks that machines are able to execute. Machines learn these tasks with the help of machine learning and deep learning for specific cases. Together, these three processes are utilized in data science so that data can be converted into insights. Since unique insights usually translate into profit for businesses, it is safe to assume that we will continue to see more and more of these autonomous machine processes in the future!

The Future..