Published in  
u
#AIunknownunknowns

Unknowns

The biggest revelation that one can achieve when analysing data is to come across ‘unknown, unknowns’, that which we are unaware of its existence. It is the one thing that makes the further development of machine intelligence worth our efforts, above the virtues of automation and robotics.

As emotionally intelligent beings, we have spent centuries pondering about the origins of our world: from the existence of a creator of all that exists, to the meaning of life itself. We, neotenic beings driven by the powerful combination of restless curiosity and versatile intelligence, have built our civilisation on the mason stone of ‘revealing the unknown’ as a way to create scientific progress, awareness of our purpose in the universe, and the humble acknowledgement that, deep down, “we know that we know nothing at all” as Socrates concluded. This is what keeps us moving forward as scientific explorers, philosophers, business entrepreneurs, social activists, parents of children, artists and creative workers. 

I was in my early twenties, I was learning, but I was curious and relentless. I just loved to decipher anomalies

Machine learning can reveal ‘unknown, unknowns’ via mathematical algorithms that we test and run on the basis of our own hypothetical guesses and assumptions. Sometimes, odd and ‘far out’ results that emerge unexpectedly force us to reverse-engineer our assumptions in order to understand the realities that they exhibit. My most treasured experience of this was in 2016, when working in Rio de Janeiro on a data analytics project. By way of an off-piste challenge, I accepted the dare that a marketing manager at Nespresso Brazil put on our desks: to reveal why Brazilians were not buying their flavoured pods of caffeine magic at the levels that they were forecasting for a market where they had been operating for ten years. When the first results came out, we confirmed what they could derive from their sales figures: Brazilians hated the limited editions. By why? And here is where the mother of all FMCG unknown, unknowns was revealed: because the concept of happiness in the Brazilian psyche is radically different to what we understand in the Northern Hemisphere. Brazilians perceive happiness as the natural state of being of every human. We are born happy and no one or nothing will truly make us happy because that feeling is within our DNA and does not come from outside inputs. So limiting the enjoyment of something is not only cruel but a waste of time when we are already happy. Contrarily, we Northerners believe that happiness is something that we must attain, win, attract as an object, that people and things make us happy. So five minutes of something is at least a brief moment of bliss in our miserable existence. How we got to think like this I shall leave to another moment, but to finally understand something so deeply cultural became an unknown, unknown that helped us strategise how to position any product in Brazil from thereafter. I always tell my data teams that the truth is out there and that we have to reveal it, that nothing exists without purpose, meaning or consequence. What data reveals is not gratuitous, if the algorithms are not biased and end up overfitting, that is, presenting as true complete nonsense information.


No items found.
The ethics wave of change has arrived at the shores of the one science that deals with living creatures other than humans all thanks to intelligent machines.

The pursue of unknown,unknowns has driven my career ever since I came across data on the trading floor of Goldman, Sachs & Co. in my training programme of 1994-95. I was rotating desk to desk on the Equities floor and landed on the emerging markets secondary market trading desk. Brazil, compared to the rest of the LatAm exchanges looked like an utter carnival of guesses every Monday morning. Perfectly sound shares were dumped against fundamentals to later on Tuesday or at end of trading, be purchased again. I asked why. I was told it was the usual in a mad market, in a chaotic country, and the way to get used to an index that took an hour break for lunch (yes.believe.you.me - the Bloomberg trading graphs would draw down a vertical line which would gap the trading for an hour each day). I firmly believed something was on. I sought data, information, assumptions, conversations with Bovespa floor traders. I found my unknown,unknown: Monday morning trading was directly affected by the Brazilian football league results. If any of the five (and there are now seven!) Sao Paolo football teams lost their weekend games in the series A, B or C, the local traders would react to it like Pavlov dogs: euphoric, thus buying like it was Black Friday;  apathetic, then dumping shares like they were worthless. I was in my early twenties, I was learning, but I was curious and relentless. I just loved to decipher anomalies. Lee Vance the Partner heading Derivatives came over the desk to meet me. I could not explain how I did it. He told me keep up the good work. And I did. All the way to today.


On February 25th in London I did a keynote for Kisaco Research at their Veterinary Health European Summit. I talked about A.I. in Animal Healthcare and I brought up the most ground-breaking ‘unknown, unknown’ in this sector: how to reveal when animals are in chronic pain, not sudden pain as in, ‘you stepped on my tail!’, but constant, unmitigated pain. Animals have no facial expressions, or do they? The answer is that they do. Not to our eyes, those easy to deceive, age deteriorating eyes that magicians trick in wondrous ways, but to the eyes of intelligent machines. A.I.-driven image diagnosis, something that in the medical world we have been developing for a decade, finally made in-roads into veterinary. Why this is ground-breaking now is that, for the very first time, the pain of animals is looked into with better tools, that the care and prevention of diseases in animals is addressed via algorithms and neural networks. The ethics wave of change has arrived at the shores of the one science that deals with living creatures other than humans all thanks to intelligent machines. The unknown revealed have ignited a new veterinary care which the vets, farmers and pet owners are benefiting from, let alone the animals themselves.

To finally understand something so deeply cultural became an unknown, unknown that helped us strategise how to position any product in Brazil from thereafter.


Socrates and his paradoxes - because statements that he made seemed to defy the common sense of his times, seem to be getting traction some twenty-two centuries later. That ethics and technology align with each other to create our future, is what he referred to as ‘virtue’ (thought, sense, judgement, practical wisdom, prudence) aiding ‘knowledge’, our desire to reveal the truth, which is what we scientists chase for the highest benefit of all concerned. 


In this decade, the revelation of unknown, unknowns will not only innovate sectors, transform geopolitics, and ignite new economic wealth, but elevate our desire to build a fairer, more caring society, kinder, more encompassing, diverse and inclusive, and this is a new sign of times emerging from the ashes of the twentieth century mindsets and nineteenth century business models that we should crash and burn. #AIunknownunknowns