Predicting the future with the past
We are creatures of Habit
Can you predict the future by looking at past data? Well, mostly yes! But it will never be 100% true. The reason why we can predict the future from the past is that we all are creatures of habit. If we tend to drink coffee in the morning, we will keep buying coffee, and that too if we have a liking towards a particular brand, we will keep continuing to buy the same brand. Some of us may be adventurous to switch to other brands, but the majority don’t. As long as we are in our comfort zone, we tend to repeat our patterns. A drastic change doesn’t happen now and then. Maybe once in a blue moon. If we travel 2 weeks in a year religiously, our destination may change, but we are likely to keep up the trend. If we and our loved ones have no health issues and as long as the money in the bank is on an upward trajectory, we don’t feel the need to change. There is an emotional lacuna that our needs always want to fill.
Silos of personal information
Now, this is exactly the kind of information that computers yearn for. Information about our habits, our daily routines, our preferences, and just about anything that they can pin down as being us. What we do at home and in our offices are nothing but events. What matters is not just the events but our reaction to them. Now if these events are captured for say the last 10 years, a computer can picture your portrait pretty much accurately. There is just one hitch. All this data is not in one place. Say you buy a new dress from Amazon; the data is with Amazon. Say you communicate with your colleagues on Slack, the data is with Slack. Your mobile phone call information is with your mobile service provider. Your wealth information is with your bank(s). And so on. Now each of these service providers has specific information about you and they all create your persona from the available data.
Auto ML and then what?
Imagine that all this information is collated and transferred into a robot. Will that robot, be you? We are still many years away from such a scenario. But technology is progressing fast. The algorithms that AI uses are becoming more and more sophisticated. Today a data scientist or an AI expert doesn’t spend too much time on modeling, because Auto ML is available today. Auto ML means that the computer would recommend the best AI algorithm depending on the data that you feed it. (e.g.: H2O, Azure ML services, etc.) Now the only thing that we need to do is to validate our data. This takes up most of the time that a data engineer or an AI engineer spends time with. Tools for doing these activities are also on the rise. The ease with which data can be munged and wrangled will become a cornerstone of the future. Not just structured and semi-structured types of data. Even unstructured data like video and voice are being annotated on the go. This is because of the availability of powerful hardware and software.
The future belongs to prescriptive Analytics
This brings us back to the question – how good are these predictive analytics? We are in this wave right now. The next big thing will be prescriptive analytics, which is being touted by many companies and experts. Prescriptive analytics means the ability of the machine to give you a solution to the problem that you are facing. Not very far from now. As data from various places gets comingled the machine has a better understanding of you. Your thoughts and emotions can also be recorded with nearly 70-80% accuracy today. Is there a hidden threat that we are not aware of? Well, this is the other side of privacy. Do we want to share everything about us with the machine(s) and trust them with our data? The vote is divided. As long as the data that we share is used for our progress, it’s agreeable. But if the data falls into the hands of a bad actor or a biased algorithm, we will have to raise our defenses. Ultimately, what we will be left with (once machines have all the info about us) will be our Spirit, which no machine will ever be able to get insights into.
And remember,
Infinity cannot be contained, no matter how powerful the machines become.
God Bless!
