Monthly Newsletter | 1st November 2016


  • Current Article
  • News
  • Quote
  • Previous Issue
  • Facebook Posts
  • Contact us
Write to Us at

Decoding Artificial Intelligence, Machine Learning, Deep Learning

Artificial Intelligence, Machine Learning, Deep Learning are currently some of the buzzwords in the technological landscape. Every time a major milestone is achieved in the field of cognition, these terms endorse the fascinating things AI can achieve. As an instance, when Google DeepMind's AlphaGo defeated the Go champion Lee Sedol earlier this year, the terms AI, machine learning, and deep learning were used in the media to describe this incredible feat achieved by DeepMind. In fact, all three are in certain sense parts of the reason why AlphaGo defeated Lee Sedol. But are they the same things? Certainly not.

The most simple approach to think about their relationship is to imagine them as concentric circles with AI — the idea that started things out — the biggest, then machine learning — which bloomed later, lastly deep learning — which is driving today's AI boom — fitting inside both.

Artificial Intelligence -
Artificial Intelligence is a branch of Computer Science concerned with empowering Machines with intelligence. The field is quite old with the term Artificial Intelligence coined in 1956 by John McCarthy at the Massachusetts Institute of Technology.

The earlier pioneers in AI dreamt to create complex machines that possessed characteristics similar to human beings. This is the concept of "General AI" (something similar to that shown in sci-fi movies like "The Terminator"). But it was sidelined later due to the technical and infrastructural challenges.

In a broader sense, the central problems of AI research includes reasoning, knowledge representation, planning, decision making, learning, natural language processing, perception, vision, the ability to move and manipulate objects (Robotics), etc.

AI tries to solve these problems using approaches like statistical methods, computational intelligence, Symbolic AI, Mathematical Morphology, Soft Computing (eg. Machine Learning), etc.

What is done currently in the field of AI falls into a concept called "Narrow AI". Basically, these are technologies that perform a specific task better than humans can. To put it in simple terms, the General AI problem is broken down to small, feasible pieces like computer vision, speech, natural language processing, reasoning, decision making, etc. Examples of narrow AI are Face Recognition on Facebook, Image classification on Pinterest.

These Narrow AI technologies exhibit some facets of Human Intelligence and they do so using one of the approaches mentioned above. This gets us to the next concept of our discussion, Machine Learning, one of the many approaches to achieve AI.

Machine Learning - An Approach to achieve AI
Machine learning is the subfield of computer science that "gives computers the ability to learn without being explicitly programmed". To put in simple words, it is the practice of using algorithms to understand data, learn from it, and then make a determination or prediction about something in the real world. So rather than hand-coding software functions with a specific set of instructions to accomplish a particular task, the machine is “trained” using large amounts of data and algorithms. This gives it the ability to learn how to perform a particular task.

Machine Learning was pursued by some researchers in the early days of AI. They were interested in having machines learn from Data. Their algorithmic approaches included clustering, inductive logic programming, reinforcement learning, decision tree learning and Bayesian networks among others. But none of them achieved the ultimate goal of General AI, and even Narrow AI was difficult using early machine Learning approaches.

One of the very best application areas for ML for many years has been Computer Vision. But the algorithms required a great deal of hand-coding like coding classifiers like edge detection filters; shape detection to determine if a shape had eight sides etc. Also, these algorithms were highly unoptimized and were further hurdled by the unavailability of high power GPU machines in the initial days.

As the field of Machine learning matured, new algorithms and techniques emerged and the right learning algorithms made all the difference. This gets us to the next concept of our discussion, Deep Learning — An Approach for Implementing Machine Learning.

Deep Learning — An Approach for Implementing Machine Learning
An algorithmic approach from early pioneers in machine learning was Artificial Neural Networks which remained unnoticed for decades.

Neural Networks were inspired by the study and understanding of human brain. Neural Networks (also referred to as connectionist systems) are a computational approach which is based on a large collection of neural units loosely modeling the way the brain solves problems with large clusters of biological neurons connected by axons.

But, unlike a biological brain where any neuron can connect to any other neuron within a certain physical distance, these artificial neural networks have discrete layers, connections, and directions of data propagation.

You might, for example, take an image, chop it up into a bunch of tiles that are given as input to the first layer of the neural network. In the first layer, individual neurons take in the data and pass it onto the second layer. The second layer of neurons does its task and passes it further, and so on, until the final layer, and the final output is produced. Multiple processing layers are involved and hence the term “Deep Learning“.

Today, image recognition by machines trained via deep learning in some scenarios is better than humans, and that ranges from identifying cats, faces, people to identifying indicators for cancer in blood and tumors in MRI scans.

Google’s AlphaGo learned the game and trained for its Go match. It tuned its neural network by playing against itself over and over and over, till it mastered itself enough to beat the Human Go champion. Thanks to Deep Learning, AI has reclaimed its charm and has a Bright Future.

Deep Learning has enabled many practical applications of Machine Learning, and by extension the overall field of AI. Deep Learning breaks down tasks in ways that make all kinds of machine assists seem possible, even likely. Driverless cars, better preventive healthcare, better product recommendations, are all here today or on the horizon. AI is the present and the future. With Deep Learning’s help, AI may even get to that science fiction state we’ve so long imagined.

To summarize, Artificial Intelligence is a broader terminology aiming to empower Machines with intelligence. Machine Learning is an approach to achieve AI, and Deep Learning is an approach for implementing Machine Learning.

Hope this article clears your understanding of the relation between Artificial Intelligence, Machine Learning, and Deep Learning.


Elon Musk’s Plan to Colonize Mars
Musk explained four key issues that must be addressed to make colonization of Mars possible: the rockets need to be fully reusable, they need to be able to refuel in orbit, there must be a way to harness energy on Mars, and we must figure out more efficient ways of traveling.
Machine Learning Tools Help Google Science Fair Finalists Find Lost Objects, Predict Breast Cancer Risk
A Silicon Valley girl from Cupertino, Calif., Cheerla was curious about the current state of breast cancer prediction, and discovered that prediction methods using digital mammograms are just 64 percent effective, typically simply considering the percentage of dense tissue in a breast.
India retains its position as world’s third largest startup base
According to the report, the number of tech startups in India is expected to grow by 10-12% to over 4,750 startups by the end of 2016. Interestingly, over 1,400 new startups emerged in 2016 denoting that that the ecosystem is becoming prudent with both investors and startup founders focusing on profitability and optimising the overall spend. With this impetus, India will become home to over 10,500 startups by 2020, employing over 210,000 people reveals the report.

““Why give a robot an order to obey orders—why aren't the original orders enough? Why command a robot not to do harm—wouldn't it be easier never to command it to do harm in the first place? Does the universe contain a mysterious force pulling entities toward malevolence, so that a positronic brain must be programmed to withstand it? Do intelligent beings inevitably develop an attitude problem? (…) Now that computers really have become smarter and more powerful, the anxiety has waned. Today's ubiquitous, networked computers have an unprecedented ability to do mischief should they ever go to the bad. But the only mayhem comes from unpredictable chaos or from human malice in the form of viruses. We no longer worry about electronic serial killers or subversive silicon cabals because we are beginning to appreciate that malevolence—like vision, motor coordination, and common sense—does not come free with computation but has to be programmed in. (…) Aggression, like every other part of human behavior we take for granted, is a challenging engineering problem!””

― Steven Pinker (A psychologist, cognitive scientist, and author of popular science)

Previous Issues

Genetic Algorithms: An Introduction With Example

Dated: 3 October 2016

Social Networking and Artificial Intelligence

Dated: 1 September 2016

A World of Chatbots

Dated: 1 August 2016

Spreading Awareness on the Importance of Technology Vision

Dated: 1 July 2016

While you were away
Artificial intelligence that spots abuse and harassment could be the answer to internet trolls
Jigsaw, an organization that once existed as Google’s think tank, has now taken on a new life of its own and has been tasked with using technology to address a range of geopolitical issues.
The latest software to come out of the group is an artificial intelligence tool known as Conversation AI.
These are three of the biggest problems facing today's AI
While companies like Google are confidently pronouncing that we live in an "AI-first age," with machine learning breaking new ground in areas like speech and image recognition, those at the front lines of AI research are keen to point out that there’s still a lot of work to be done. Just because we have digital assistants that sound like the talking computers in movies doesn’t mean we’re much closer to creating true artificial intelligence.
There is always and will be a word-war over Natural Stupidity and Artificial Intelligence.
Let's see what a theoretical physicist Stephen Hawking and a business magnet & inventor Elon Musk has to say over this word-war.

The Centre for Existential Risk features classes that deal specifically with potential problems such as war and climate changes that may face humanity in the not-so-distant future. It’s about time, too. According to a recent press release by the CER, artificial intelligence that rivals or even surpasses human brainpower may be developed within the century.
Contact Us

Your reviews are valuable. Help us to improve.
Give your feedback on

Address: 10, Jamanadas Industrial Estate, Phase -1, Dr. R.P. Road, Mulund (West),, Mumbai, Maharashtra 400080

© Copyright Artificial Intelligence Mumbai. All Rights Reserved.