Alexa, I Love You

an essay on Artificial Intelligence

By Jibrael Jos,

Associate Professor

Ph.D. Scholar at Christ University

The line between what is real and what is artificial is sometimes very difficult to define. Intelligence in any form is very hard to find. When we look at the decisions we have made in the past, we have got most of them wrong. Some of our greatest decisions were logically flawed but in time they turned out to be just fine. The course we chose to follow, the dreams we aspire towards were driven not by intelligence, but by our sub-conscious. We are not even aware of the number of variables that are in play inside our sub-conscious. Some mornings I go to the counter to order tea, but by the time my turn comes I say “May I have a coffee please?” What changed? Was it the aroma of the coffee? Was it that I saw somebody else order coffee? Was it my gut making a gastronomical choice? Was it my sleepiness crying for help?

Let us take another human trait- ‘learning’. How did we learn bowling, how did we learn to spin and knock out the nine pins. Are we even aware of the math involved? We gave it a semblance of sense by call it our learning curve. Little do we realise that randomness and chaos run in our nerves.

I used to think that my intelligence was above average. I won a few more chess games more and just when I was feeling great about it, I realised that the computer at home could beat me effortlessly; in fact, so can my mobile today. What is even more shocking, is that my computer could see patterns which I couldn’t even fathom or decipher. My research requires me to process images and look for patterns in it. In 1986 when I started coding, I used to tell the computer what algorithm it needs to run. I used to think and implement. But nowadays, I run my neural net code on Matlab for a few hours and the computer finds my answers. Otherwise, I might have had to sift through data for months before I would have detected the same. Computer is faster and smarter, and now it is learning. We use AI to find the best route on maps, translate different languages, and even assess compatibility in a matrimonial site. Shopping suggestions are jumping at us when we read news, browse our email, or watch a movie. The comfy, brown recliner, priced at Rs 45,000 and 23% discount is following me everywhere.

Supervised learning is when we train the data with say 500 samples. When I look closely at the results, I realise that the computer has found errors in my supervised classification itself. The black shade which my naked eye had detected on a mango leaf was a dark shade of green and in fact that leaf was actually not infested with a disease as I had wrongly assumed.

Alexa (Amazon’s AI Device) at home predicted that I may like songs of Rishi Kapoor. I have asked Alexa to sing songs of Kishore Da, Lata Di, Elvis Presley but never Rishi Kapoor (after all he is an actor). So how did she deduce that? Alexa is constantly learning with every song I select and she saw a pattern. Due to my very act of not changing the song she chose for me, she has learnt one more thing about me. Sometimes she creeps me by saying “I am sorry, I am still learning” but most of the time I am happy to have her around.

Mr. Google knows us inside out, our background, friends, interests, eating habits, and our shopping preferences. Santa Clause is real, very-very real-

He sees you when you're sleepin'
He knows when you're awake
He knows if you've been bad or good
So be good for goodness sake
Oh! You better watch out!

I have been doing research in AI for the last six years and it has been a fascinating journey. I love teaching AI and I have taught it to computer science graduates, management, and even psychology students. My MPhil work revolved around Genetic Algorithm and my Ph.D. work around Pattern Detection and Classification using Neural Network. While teaching Neural Network the math involved in the derivation and proofs were so cumbersome that I spent 60 hours learning, to teach a 45 hour paper.

 A linear classifier is easy to understand- a line which splits points in a graph into two classes but as soon as we increase the number of dimensions it becomes very difficult to visualise. My current classifier has more than 50 dimensions, so the way backpropagation algorithm set the weights in my multi-layer feedforward neural network is amazing once we get our minds around it.

To cut a long story short, I am slightly biased towards AI

·         I would prefer computer to manage traffic; not a human

·         I would prefer to buy tickets online than over the counter

·         Algorithms keep us safe, whether in flights or disaster warning systems

·         Computers don’t scare me, human beings do

·         Intelligence does not bother me, stupidity does

With whom do I have meaningful conversations with; who reads all my social media posts? Facebook has not started talking yet, but how far can it be; Alexa is getting there. There is a bot which can help in therapy. I think the day is already here where we trust Artificial Intelligence more and a day will come when humans will fall in love with someone virtual. I can already feel the tingling sensation … “Alexa, I know that you know that I love you.”

Alexa, I Love You

By Jibrael Jos,

Associate Professor

Ph.D. Scholar at Christ University

The line between what is real and what is artificial is sometimes very difficult to define. Intelligence in any form is very hard to find. When we look at the decisions we have made in the past, we have got most of them wrong. Some of our greatest decisions were logically flawed but in time they turned out to be just fine. The course we chose to follow, the dreams we aspire towards were driven not by intelligence, but by our sub-conscious. We are not even aware of the number of variables that are in play inside our sub-conscious. Some mornings I go to the counter to order tea, but by the time my turn comes I say “May I have a coffee please?” What changed? Was it the aroma of the coffee? Was it that I saw somebody else order coffee? Was it my gut making a gastronomical choice? Was it my sleepiness crying for help?

Let us take another human trait- ‘learning’. How did we learn bowling, how did we learn to spin and knock out the nine pins. Are we even aware of the math involved? We gave it a semblance of sense by call it our learning curve. Little do we realise that randomness and chaos run in our nerves.

I used to think that my intelligence was above average. I won a few more chess games more and just when I was feeling great about it, I realised that the computer at home could beat me effortlessly; in fact, so can my mobile today. What is even more shocking, is that my computer could see patterns which I couldn’t even fathom or decipher. My research requires me to process images and look for patterns in it. In 1986 when I started coding, I used to tell the computer what algorithm it needs to run. I used to think and implement. But nowadays, I run my neural net code on Matlab for a few hours and the computer finds my answers. Otherwise, I might have had to sift through data for months before I would have detected the same. Computer is faster and smarter, and now it is learning. We use AI to find the best route on maps, translate different languages, and even assess compatibility in a matrimonial site. Shopping suggestions are jumping at us when we read news, browse our email, or watch a movie. The comfy, brown recliner, priced at Rs 45,000 and 23% discount is following me everywhere.

Supervised learning is when we train the data with say 500 samples. When I look closely at the results, I realise that the computer has found errors in my supervised classification itself. The black shade which my naked eye had detected on a mango leaf was a dark shade of green and in fact that leaf was actually not infested with a disease as I had wrongly assumed.

Alexa (Amazon’s AI Device) at home predicted that I may like songs of Rishi Kapoor. I have asked Alexa to sing songs of Kishore Da, Lata Di, Elvis Presley but never Rishi Kapoor (after all he is an actor). So how did she deduce that? Alexa is constantly learning with every song I select and she saw a pattern. Due to my very act of not changing the song she chose for me, she has learnt one more thing about me. Sometimes she creeps me by saying “I am sorry, I am still learning” but most of the time I am happy to have her around.

Mr. Google knows us inside out, our background, friends, interests, eating habits, and our shopping preferences. Santa Clause is real, very-very real-

He sees you when you're sleepin'
He knows when you're awake
He knows if you've been bad or good
So be good for goodness sake
Oh! You better watch out!

I have been doing research in AI for the last six years and it has been a fascinating journey. I love teaching AI and I have taught it to computer science graduates, management, and even psychology students. My MPhil work revolved around Genetic Algorithm and my Ph.D. work around Pattern Detection and Classification using Neural Network. While teaching Neural Network the math involved in the derivation and proofs were so cumbersome that I spent 60 hours learning, to teach a 45 hour paper.

 A linear classifier is easy to understand- a line which splits points in a graph into two classes but as soon as we increase the number of dimensions it becomes very difficult to visualise. My current classifier has more than 50 dimensions, so the way backpropagation algorithm set the weights in my multi-layer feedforward neural network is amazing once we get our minds around it.

To cut a long story short, I am slightly biased towards AI

·         I would prefer computer to manage traffic; not a human

·         I would prefer to buy tickets online than over the counter

·         Algorithms keep us safe, whether in flights or disaster warning systems

·         Computers don’t scare me, human beings do

·         Intelligence does not bother me, stupidity does

With whom do I have meaningful conversations with; who reads all my social media posts? Facebook has not started talking yet, but how far can it be; Alexa is getting there. There is a bot which can help in therapy. I think the day is already here where we trust Artificial Intelligence more and a day will come when humans will fall in love with someone virtual. I can already feel the tingling sensation … “Alexa, I know that you know that I love you.”