Thursday 30 August 2018

Sophia and Me - Artificial Intelligence, Machine Learning, Deep Learning


It was 7.30 pm and raining heavily.  I stood in front of the Seminar hall  waiting for my Uber. 

I checked my smartphone .  

My cab was just about 1 minute away.  It arrived and I swiftly ran from the shade of the outer atrium of the hall towards the gate, got into the taxi and heaved a sigh of relief.  

But my relief was very short lived. 

“Let’s go !”  I said.

The Uber driver was fiddling with his phone - shaking  wiping , putting it near his cheeks and ears .

“Sorry Mam, I cannot go. Please cancel this ride and get another one “,  the driver said .

“What ! But Why ?  It is raining so heavily , can’t you see ?  And I have a doctor’s appointment at 8 pm ! “

“Mam, I can go but you will have to tell me the directions.  GPS is not working because phone is damaged. It fell into a puddle when I was helping the earlier passenger with her bags.”

“Oh God !” I groaned. 

I tried to open the net but by that time it has started raining cats and dogs. My current location was below a flyover and there was almost no network. And the torrential rain made the connectivity almost nil.

I was really flabbergasted. 

I was going to this doctor for the first time. In the morning , my husband had asked me to look up the direction on google maps but I had refused.

“I will anyways get an Uber . With technology , I don’t even have to communicate with the chauffeur. I will reach on time. Don’t you worry !” I had laughed.

I never imagined that my laugh will cost me so heavily !  

And at that moment, I truly realized the extent of our  dependency on technology.  

Or you can call it on Artificial Intelligence.

Artificial Intelligence, Machine Learning and now the most recent , Deep Learning

Let me first try to explain these in layman terms.

Artificial Intelligence is the broad umbrella term for attempting to make computers think the way humans think, be able to simulate the kinds of things that humans do and finally the machines  solve problems in a better and faster way than we do. 

John McCarthy, widely recognized as one of the godfathers of Artificial Intelligence (AI), defined AI as “the science and engineering of making intelligent machines that have the ability to achieve goals like humans do. In short, Artificial Intelligence is human intelligence exhibited by Machines.

In fact AI has been around for quite a while. Though the term  Artificial Intelligence was coined at the Dartmouth Conference in 1955, simple things like computers, calculators, computer games , washing machine, dishwasher – anything which the machine does faster and better than human beings comes under the umbrella of Artificial Intelligence.  

It was only since 2015, the domain of AI  started becoming more focused and narrow. 

This gave birth to Machine  Learning . The idea behind Machine Learning is fairly simple.

Rather than programming computers to be smart by hand-coding software routines with a specific set of instructions to accomplish a particular task, you give machines access to a large number of sample data and code them to find patterns and learn on their own how to perform the task. 

Machine Learning is built based on algorithmic approaches that over the years included decision tree learning, inductive logic programming, clustering, reinforcement learning, Bayesian networks etc.

Let me try to make it  simpler. 

When I taught the English Alphabets to my daughter - ,A for Apple, B for Ball, C for Cat,   I did not tell her the algorithm to identify the features and then decide what is it. I simply showed her the picture of an apple , a ball  and a cat.  When I bought fruits , I showed her apples - some green , some red. When I went to the park as saw a cat sitting there, I held her hands and  said “ See , a Cat !” . 

So when she saw multiple examples of the  object and then her human brain automatically identified the features so that her brain could identify that object

This is  exactly what a Machine Learning Model does. Human brain is a machine par excellence. The Same thing is simulated in Machine Learning.  Huge amount of data is fed into the computer and after analysis of the data, prediction can be made and some action is taken.  

Google uses Machine Learning to filter out spam messages.  Amazon and other top technology driven e-commerce sites use machine learning wand data analytics to advertise their goods based on what the algorithm predicts about your needs and tastes.

Now about  Deep Learning. This is even more focused and intense , the next level of AI.

Deep learning was inspired by the structure and function of the brain, namely the interconnecting of many neurons. Neural Networks are algorithms that mimic the biological structure of the brain. In the brain, the neural points are physically spaced out, but in Deep learning, there are layers of neural network. 

Deep Learning can be explained as a system of probability. Based on a large dataset you feed to it, it is able to make statements, decisions or predictions with a degree of certainty. So the system might be 63% confident that there is a cat on the image, 81% confident that it’s an animal and 4% confident it’s a toy. Then you can add on the top of it a feedback loop, telling the machine whether it decisions were correct. That enables learning and possibility to modify decisions it takes in the future.
Again I will try to make it simpler.  
Suppose there is a traffic signal detector installed in a driver less car . After processing billions and trillions of data, the first layer will decide that there is a probability of 81 % that there is a rectangle . This  decision will passed on to the next layer of the neural network which will again process a humongous amount of data and give a signal to the next layer that there is a 72% probability that the rectangle is a traffic signal with red , yellow and green light on it. This then goes to the next layer and to the next and so on.  

Finally the signal detector will take a decision whether to drive ahead or not. And  it will take back data cues whether the decion was correct or not and improve itself for doing better next time. 

As in my previous example of teaching alphabets to my daughter, Deep Learning can be compared to the  way she became capable of learning from mistakes and constantly improving. Like when she called  the dog a cat  and she got a poor grade in her test,  she was given the feedback, she corrected herself and improved . She then became capable of learning new objects on her own.   
Deep learning is being implemented for driver less cars, face recognition,  pattern recognition, forecasting, the list is huge.  
The future is wide.  But somehow I tremble at the thought of the future.  

When I start thinking , I am scared. 

Will machines replace human beings ? Already we are at a stage where for several tasks, we trust the machine more.  Earlier, for giving loans, the data was analysed by human beings for deciding whether to give the loan or not. But nowadays, the banks rely on the computer algorithm , which scans and processes huge amount of data and tells us whether to process the loan or not.  

If you ask the bank why your loan was not sanctioned, they might just shrug their shoulders and say – I don’t know but the computer said so ! 

 And like by Uber driver refused to complete my ride because of a technological glitch,  will a time come when machines will literally rule us ?

In the Industrial Revolution when F.W Taylor came up with the scientific model of management,  it gave rise to processes and greater productivity. But assembly line production in factories, created jobs too. 
But with Artificial Intelligence,  the human jobs might be completely eliminated
And it's not just in factories, but truckers, drivers and even jobs like telesales, customer service and hematologists as well as radiologists over the next 15 years are going to be gradually replaced by artificial intelligence.
 Sundar Pichai, the CEO of Google has said that “in the next 10 years it will be AI first !” .
AI is cheaper and more productive . It doesn’t sleep, need breaks, get sick or take vacations, and it doesn’t need health insurance or retirement benefits. It can work 24X7, faster, better and accurately. With Deep Learning, AI can  quickly acquire and learn new skills too  without making mistakes ! 
Isn’t this too scary ?

I know that repetitive and boring jobs will be replaced by AI. But many learned people have predicted that by 2050, there will a big chunk of people who will not only be unemployed but also unemployable !  
So what will happen to my economically poor housemaid,  my not so intelligent, personal driver  and the good natured neighborhood carpenter ? How will they sustain themselves ? 
AlphaGo is a computer program that plays board game.  It has been developed by Google’s DeepMind, using Deep Learning.  
Reently AlphaGo defeated the world champion Ke Jie.   

And how Ke Jie cried ! 
With all these technological advances, have we become happier than our forefathers ?  I really don't think so ! 
We have already lost simple joys of life like receiving letters  from our dear ones and preserving them for later. Many a times I have taken out the  bundle of letters written by my husband during our courtship days , read and re read them , feeling the written word, smelling the musty smell of  the paper. 
The  joy of waiting for my Uncle coming to visit us, running towards the gate whenever I heard the sweet tinkle of the bell of the rickshaw , cursing the train service for being late and then jumping with joy when he scooped me in his arms when he finally came home
Simple thing like peeking out of the taxi window to ask a shopkeeper or a passerby about an address location and smiling back at him.
In these interesting times, will a time come when humanoids will exist together with human beings ? 
It is already a reality.
Sofia, the first humanoid robot , made her first public appearance in 2016. 
She is a doe eyed woman with delicate features, designed to look like Audrey Hepburn. 
AI allows her to hold eye contact, recognize faces and understand human speech. Cameras within Sophia's eyes combined with computer algorithms allow her to see. She can follow faces, sustain eye contact, and recognize individuals. She is able to process speech and have conversations using a natural language subsystem. She has become a citizen of Saudi Arabia  !
Ke Jie cried when he lost the game to AlphaGo. He loved his game so much that it broke his heart.  
Did AlphaGo feel  pleasure or happiness when it won ?  Definitely not .  
And Sophia, she might become super intelligent but her deadpan voice tone, mechanical facial expression will never be able to replicate the rush of Adrenalin and the heavenly emotions of love, pain and joy.
We are  living in very interesting times.  Let us see what holds in the future. Maybe it will be like what Charles Dickens had  expressed  in the immortal opening lines of  his novel - The Tale of Two cities  
It was the best of times, it was the worst of times, it was the age of wisdom, it was the age of foolishness, it was the epoch of belief, it was the epoch of incredulity, it was the season of Light, it was the season of Darkness, it was the spring of hope…..
Let us hope for the best !

Saturday 25 August 2018

5 things to avoid while speaking in Impromptu Situations


If you want to speak as a leader in impromptu situations, avoid the following blunders:
1. Talking Too Much
A common pitfall is blabbering  on. Those who have this tendency, interrupt others, monopolize meetings, ignore valuable contributions of others, and in general turn people off.
Stick to the matter at hand. Don’t waste anyone’s time.
2. Not Preparing
Winston Churchill had fun with speakers who talk without thinking. He observed: “Before they get up, they do not know what they are going to say; when they are speaking, they do not know what they are saying; and when they have sat down, they do not know what they have said.”
Well, we can prepare to be spontaneous. There are times when you know you’ll be part of an impromptu event – a client chat, a conversation with a subordinate, an important visitor. Take whatever time you have to prepare notes, or a mental outline of what you will say. Even if you only have a few seconds, pause and decide what your message is.
3. Misreading your Audience
Still another impromptu faux pas  is misreading your audience.
Be sensitive to the feelings, concerns, and knowledge of your listeners. Some of the engineers I’ve coached are spot on when speaking , but overestimate the ability of colleagues and clients to understand particular issues.
Similarly, a boss might misread a situation in which a new hire needs a compliment, or friendly guidance, rather than a critique. A participant in a meeting may speak up with a proposed plan, not realizing (because he has tuned out) that the group has just moved beyond that proposal. Staying attuned to your audience’s thinking is critical in all situations.
4. Letting off Steam
Another impromptu gaffe is letting off steam. We all carry baggage with us\ in our minds. Fears, disappointments, concerns, and uncertainties. Occasionally impromptu exchanges bring these to the surface. As a result, sometimes we say things we regret.
5. Cracking Jokes
Humor is dangerous for leaders if not properly thought of !
An executive I know was honoring a staff member, who was about to retire. He stood up and began his speech: “Sunny is a memorable figure. One colleague will remember him for being late…. another for his crazy sense of humor…. and still another for the fact that he has always been a bit wierd.” He concluded by saying, “So we’ll always remember you, Sunny, for the things you may wish to forget!”
Everyone laughed – Sunny the loudest – but no one was laughing inside. They were embarrassed for Sunny and for the speaker. This was not a moment of leadership.
All these pitfalls can be avoided if you prepare for your impromptu remarks.
The underlying message of Impromptu is that extemporaneous speaking requires forethought, discipline, and preparation.  Keep key leadership messages uppermost in your mind. And use whatever time you have to read your audience, collect your thoughts, choose your words, and structure your script. These techniques will equip you for success as a leader in the Age of Impromptu.