During Google’s early days, I really liked their “I’m feeling lucky” button for two reasons. The geeks behind google search were betting that they know what is likely the best result for me. That takes guts. And on the sparsely populated white page — on the pride of technical wizards, the word “feel” found a prominent home. That pleased my right brain as well. How did they get the feel right? Here is the backstory in everyday words.
If we think of webpages as pages of the book, Google algorithm was unique — their search was like the index section in the back of the book rather than table of contents in the beginning. In a typical book, the page numbers after each word (on the index page) are arranged in ascending order of occurrence. Google reordered the page numbers after a word based on how often people referred to an internet page. The first among the results became the result of “I am feeling lucky” button. Google continued their pursuit of indexing every page on internet and showing us the time for search in centi seconds — their geeky pride.
The challenge today, the book has grown bigger and “I am feeling lucky” still exists, but as a play button like a slot machine coming up with other options. Google search, through the years, has made rapid strides on figuring out what we are literally searching based on hierarchy of rules.
Can google search morph into an educated guess on what we really hoping to search? That feels like a dream and it is.
Understanding dreams starts with the answer to this question: What does the picture look like?
Your answer, buildings architecture did not take long, did it? From your visual neurons in the eye, all the way to the processing in the folds of your brain was almost instantaneous. Can you explain step by step, how you came up with it? You may pause to figure out how you did it instantly. Intuition is a good answer. It is a sophisticated form of analysis. Something we all do so well — the beauty within us.
On the other hand, what happens if we let a machine do the same by replicating the neural connections, the synapses that create the magic within us? A deeply learning network of artificial neurons is like a house of cards (stronger and firmer than cards) with the inputs at the base and the output at the top. Every layer acts as a foundational input for the upper layer.
What happens when the machine shares its answer consistent with our intuition? The bar is higher; we would like to know how it came up with this answer. Our curiosity for logic and earnestness to verify takes over before we can outsource our trust to the machines. In other words, can we pause at every layer and get a human peek on how the black box dreams up the final result? Thanks to the rapid advances in computation speed and technology, engineers at google and labs around the world ,they are able to do just that — A sample below on how a lower layer neuron looks at the picture..
“All this is interesting, what does it mean for you and me?”- if your thoughts mirror these words, you are not alone. The simple answer- the part of the memory we have outsourced daily — google search is getting turbocharged with machine intuitiveness.
Symbols are powerful, many times, they share more than what they represent. The passing of the baton for google search happened two years ago.
The current head of google search is the former head of artificial intelligence group at Google. This was a singular moment* in our history where deep learning moved from labs to a mainstream commercial product on such a large scale.
Our Human Future: What it means for careers (for our kids next )
Feeling lucky with good educated guess on what I am searching is one thing. On the other hand, feeling a churn at the bottom our stomach is a natural reaction. The sub-text is about machines mimicking our intelligence- we are challenged in our home front. A replica of feelings (even if it is fleeting),I could relate when I first saw Arnold Schwarzenegger land on earth as the terminator in 1984!
Here is the comfort — with all the rapid advances — these deep learning machines are like horses with blinkers. They learn what they are taught in front of them. Any thing outside their realm, like connecting the dots between random topics is still the prerogative of the human mind.
Creativity is the human hallmark with our emotions ruling the roost — at least when it is time for our children’s generation to rule the workforce. That is what I want to believe.
I enjoy writing at the intersection of analytics and human relationships.
* for those who like completeness, actual tip toeing of the tool, rank brain, started in 2015 for more outlier searches.