In the world of finance, once an opportunity to arbitrage becomes popular, it ceases to exist as equilibrium is swiftly reached. The diffusion of information implies reduced profits as the pie must now be shared amongst many bidders. The most competitive arbitrageurs are the ones who discover opportunities the earliest. To us at Gauge, the world of data must be looked at through a similar lens. We believe that the key to success is to harness the hidden value within data before your competition does. Why are we at Gauge in a better position than our peers to help you do so?
We strive to be at the cutting edge. Often, when dealing with large amounts of complex data or "big data", the difference between the potency of the latest scientific technique and a slightly older variant may turn out to be phenomenal. At the very least, the latest technique might lead to a better estimate of a trend or more strikingly the extraction of an entirely new insight that might then be monetized.
Specialization is indispensable when it comes to increasing efficiency. In our opinion, the same holds true if one has to extract the most from big data. A specialized solution would most certainly trump the one that purports to do everything. Indeed it is commonplace today to find such all in one, do it yourself big data platforms. These have their place in the analytics ecosystem and may be good at making the low to middle hanging fruit more accessible. However, a specialized solution tailor made for your business would truly set you apart, allowing access to that high hanging fruit as well that potentially revolutionizes the way you do business. The best way to make sense of complex data is to combine the perspectives of many specialists from diverse fields into an actionable business insight. This multidisciplinary culture is reflected not only in the talent we hire but also the solutions we provide.
It is our belief that big data is yet to meet "big science". The state of the art in the science of big data today involves machine learning techniques that have glaring limitations. Reason being that most machine learning algorithms are of the "black box" variety. Consequently, very little may be known about how different co-factors within a data set interact to yield a certain trend or pattern. This in itself suggests that the knowledge gained is only partial as a complete understanding of the underlying causes is missing. Even more strikingly, as evidenced routinely by practitioners, it becomes impossibly hard to become more accurate beyond a certain threshold while solving "real world" problems through machine learning. This was clearly evident during the famous Netflix competition. Hundreds of teams of the best machine learning researchers working for more than a year could only best Netflix's own algorithm by a mere ten per cent. Is there life beyond machine learning? Are better insights possible via alternate tools? The answer to both these questions in our opinion is a big yes. Many of the fields that may be considered science heavyweights, including physics (particularly statistical physics), information theory, and complexity theory remain largely untapped when analyzing big data. In addition, ultimately any insight gained through data will have an interpretation component that is human. This strongly suggests that techniques drawn from the neurosciences should play a prominent role in data driven decision making. At Gauge, we are intensely focused on bringing in new science in making big data solutions that much more impactful.