April 22, 2014
Dear Fellow Owners of Health Discovery Corporation,
Recently you received information about SVM Capital LLC's ("SVMC") success in achieving a risk-adjusted tie for first place in an audited investment contest featuring quantitative techniques in which SVMC used its proprietary SVM-based algorithm. We believe this could have implications for both SVMC and Health Discovery Corporation ("HDC" or the "Company"). This development also makes for an opportune time to share our views as to where we think HDC and our "flagship" technology, the Support Vector Machine ("SVM"), fits into today's data-based technology landscape.
The world is full of impressive machines like Google glass, iPhones, robots and Boeing 767s that we can see, touch, drive, fly, work and play with. But there's another noteworthy machine, intangible and invisible, that exists only in the world of zeroes and ones but is very impressive in its own right. This is the SVM, an algorithmic invention in the field of machine learning that, in the right hands, can find computer-based solutions to problems when other mathematical or analytical efforts may fail. One of its special features is that it operates in multi-dimensional space. That is to say it can marshal many different approaches (often called kernels) to analyze and deal with limitless variables contained in huge amounts of data. The SVM, a supervised machine learning technique, is particularly useful in detecting patterns and isolating key variables in dense data thickets.
The SVM was invented by Vladimir Vapnik (who won the Humboldt Prize) and refined into more practical application by Isabelle Guyon. Both Mathematicians have provided significant contributions to HDC. In fact, it was Dr. Guyon's work for HDC that led to the discovery of genes associated with prostate cancer and a diagnostic test that HDC's commercial partner is expected to launch this summer. HDC invented some of the earliest SVM's, which are part of our 80 patents. In addition, HDC has the first patent related to SVM-RFE (recursive feature elimination) in the world. In brief, HDC has had broad experience in using SVM's in its efforts to work with partners to build commercial products.
One of the hottest topics in the hi-tech world today is "big data", a very broad term that covers a lot of waterfront. Why? Some 94% of all the world's data has been created in the past two years and continued zettabyte data flows will present incredible opportunities for many, particularly those who can analyze it. The March-April 2014 issue of Harvard magazine has an illuminating article entitled, "Why Big Data Is a Big Deal". It is well worth reading in order to get a sophisticated and comprehensive perspective. Summarizing and paraphrasing, here's why:
There is a big data revolution but it's not the quantity of data that's revolutionary; rather it's that now something can be done with it. The revolution lies in improved statistical and computational methods, not in the exponential growth of computer capacity and storage. The doubling of computer power every 18 months (Moore's law) is nothing compared to the "big algorithm" - a set of rules that can be used to solve a problem a thousand times faster than conventional computational methods could.
Easier said than done because so much data is "tangled" and multi-dimensional, that it cannot simply be unraveled by standard mathematical techniques like neural networks, Bayesian models, etc. In math jargon, when epistemic density is great, the SVM's high dimensionality can often unravel the complexity and yield a meaningful solution. At a recent conference on machine learning, an IBM expert said its approach to big data was to "cast a very wide net." But after SVM Capital LLC's architect, Mark Moore, spoke, he acknowledged to Mark that HDC's "tangled", multi-dimensional description was much more accurate and that while IBM's "big net" approach has clear benefits, it also has notable limitations.
It is not sufficient simply for work-a-day mathematicians who also "do" SVMs to try to address difficult problems for they, like musicians, brain surgeons and quarterbacks, are distributed on their own bell curve. HDC has virtuoso-level talent that can attack a problem directly and, just as importantly, train and supervise other mathematicians in SVM techniques that have already worked well for us. It is also impossible for any mathematician to solve a data problem without the intimate involvement of domain experts in their field, and this is precisely the road HDC has traveled with its partners over the years.
One of those partners is SVMC which has created an investment algorithm that in 17 months of live trading (with internal capital) has not only outperformed the S&P 500 but simultaneously reduced portfolio risk to levels below that index. This could not have been achieved without the SVM. SVMC's initial goal is to start a hedge fund investing in large and mid-cap U.S. stocks and this recent development with BattleFin suggests this is within reach. Longer term, SVMC's goal is to create comparably derived algorithms for other well-established stock markets as well as for currencies and commodities since there are similar factors influencing prices in all financial markets.
It took SVMC, 45% owned by HDC, six years to create this algorithm but that's not because the experts were slow learners. Rather, one of the Gordian Knot issues that had to be solved was using a tool (SVM) designed to deal with static data like genes and adapting it to dynamic data like stock market prices. This was not a trivial challenge and the solution required deep knowledge, imagination and a lot of patience. The end result though has been worth it. Notably, SVMC invests as distinct from "trading" as the portfolio is rebalanced only quarterly. Because of the design of the algorithm, the portfolio is hugely scalable without increasing investment complexity: the decisions required to manage $5 billion are the same as for $5 million, the only difference is that investment positions are proportionately larger.
SVMC is a "quant" fund with modest human intervention but portfolio decisions are based on rank ordering new fundamental data from Zack's on a quarterly basis. One thing that differentiates SVMC from other hedge funds is what it doesn't do.
* it doesn't "bet" on anything or use risky leverage
* it doesn't "trade" at warp speed, hourly, daily, weekly or monthly
* it doesn't chase interest rates or currencies or market trends
* it doesn't do debt or financial exotica or illiquid instruments
* it doesn't do international or pick investment sectors
* it doesn't do "sentiment" or "momentum" or "psychology"
HDC was originally founded to develop diagnostic tests, principally for cancer, by combining its mathematical SVM expertise with the intimate knowledge of "domain experts" such as M.D. Anderson Cancer Center, Quest Diagnostics, Abbott Labs, Genzyme and others. Essentially, we envisioned the "marriage of math and medicine" to create a vehicle to help develop what has come to be known as personalized medicine. More recently, IBM Medical was started with the same objective in mind.
In 2012, HDC licensed many of its patents for medical diagnostics to NeoGenomics, with whom HDC envisions a productive and profitable future. The first test is NeoSCORE™, a highly accurate blood and urine prostate cancer test that has the potential to supplement or compete with or, in time, even replace the PSA test. In addition, HDC is working diligently with NeoGenomics on Cytogenetics and Flow Cytometry projects that use the SVM to help domain experts analyze complex images.
It is almost a truism to point out that any identifiable technology today may be just an integral part of a larger paradigm tomorrow. Why? Because technologies tend to converge. For example, first there was the automobile, a technological wonder at the time and next came autos and electronics together, and now we have autos, electronics, ergonomics, new propulsion systems and computer informatics all bundled up. While experience, intuition and judgment will continue to be essential to solving problems, future decision-makers will necessarily rely more and more heavily on analytics to help with that. And this is where an ability to interpret new tsunamis of data that inundate us daily will be invaluable in simply sorting it all out, first into information and then into actionable intelligence. And in the right hands with the right domain experts, the SVM can do that exceedingly well. Almost regardless of the field or technology, HDC believes that sophisticated mathematical analytics will be a thread common to advances in all of them. There are many data-based market opportunities ripe for exploitation including medical diagnostics, genomics, proteomics, digital imaging, bioinformatics, stock market analysis, internet search and security, credit card fraud detection, health data analysis and many others.
We hope this broad-brush treatment gives you, the owners of HDC, a good idea of what your Company is and does and what we're capable of and aspire to in the many fields of "big data".
With kind regards,
The HDC Team
This document contains forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995, the accuracy of which is necessarily subject to risks and uncertainties, including, without limitation, statements regarding future performance, opportunities and investments, and anticipated results in general. From time to time the Company may make other forward-looking statements in relation to other matters, including without limitation, commercialization plans and strategic partnerships. Actual results may differ materially due to a variety of factors, including, among other things, the acceptance of our approach to applying mathematics computer science and physics into the disciplines of biology, organic chemistry and medicine and our products and technologies associated with those approaches, the ability to develop and commercialize new drugs, therapies, medical devices, or other products based on our approaches, and other factors set forth from time to time in the Company’s Securities and Exchange Commission filings.
All forward-looking statements and cautionary statements included in this document are made as of the date hereof based on information available to the Company as of the date hereof, and the Company assumes no obligation to update any forward-looking statement or cautionary statement.