Google

YouTube

Spotify

Scientific Sense Podcast

Friday, September 6, 2019

The emerging Principal-Agent-Machine problem in the enterprise

Ever since the owners of organizations put agents in charge of operations, because of growing scale and perceived need for management specialization, shareholder value maximization has not been Pareto optimal for decision-makers. Much has been written and studied in this area with little effect on organizational structure, systems, and strategies (1). From the advent of computers, agents have been effective in claiming superiority over machines because of a lack of transparency for the owners. Although it is difficult to prove that machines possess superior decision-making capabilities in real companies and markets because of the lack of data from long and repeated experiments, it is clearly the case in financial markets.

With clear and consistent data in the financial markets, it has long been clear that financial intermediaries and traders have been destroying alpha, forever. With misguided and a confusing focus on "absolute returns," these agents have been successful in siphoning out wealth from owners in fees and expenses. An illustrious investment bank seems to have recently recognized that "trading," done by humans creates no value for its clients. Machines are infinitely better as they can act based on complete information without bias. Decision-making, thus, is better delegated to machines.

In real markets, this is equally applicable. Because of high diversity in types of decisions and long durations to outcomes, agents have long claimed superior capabilities compared to machines. This is true at all levels of organizations (1) and in every function. Since distributed owners are unable to understand the inner workings of complex organizations, agents simply claimed they are better without any contention. This has significant negative effects on the economy and its potential to grow. A structural change that culminates in the reassignment of human responsibilities in the enterprise may be afoot.

The emerging principal-agent-machine problem is real for modern organizations. Institutionalization of agents since the industrial revolution has run its course. Owners may finally have an opportunity to break this stalemate.

Wednesday, August 21, 2019

Decision-making is different from finding cats and dogs

As AI moves toward the peak of the hype cycle, it is important to recognize that decision-making in an enterprise is distinctly different from training machines to differentiate between cats and dogs. Most of the field is focused on deep neural networks, convoluted and otherwise, to recognize text, pictures, audio clips and patterns. This is certainly interesting but an extrapolation from these techniques to improving decision-making in the enterprise is fraught with danger. As companies find that enterprise productivity is inversely proportional to the number of data scientists they have on staff, reality is beginning to sink in.

As technology and consulting companies try to mop up the last remaining "data scientist," on Earth, it may be interesting to take a measurement of how enterprise productivity is related to them. Data science, an ill-defined field, has been the latest hype that led many companies down rabbit holes with very little to show. Although there are interesting developments in Artificial Intelligence - in robotics and autonomous equipment, much of these are better called expert systems as they do not learn from data but work on coded heuristics. The stars in the field do not prefer the old terms such as "expert systems," and "neural networks," as they believe they have reinvented mathematics. This is symptomatic of a field beginning to go off the rails as the investors have unrealistic expectations of the "second coming," of AI.

Let's not throw out the baby with the bath water. It is just that the baby has a lot of growing up to do. Decision-making takes a lot more than supervised or unsupervised machine learning. Educational institutions do a disservice to the next generation by blindly following the latest trends and spawning "analytics courses," for everybody. The question educators should ask is whether such programs are leading to people who can take advantage of the technology to enhance enterprise value. To do this, they have to first understand how value is created and that is a lot harder than cranking the supercomputer in the cloud.

Artificial Intelligence has a lot of potential, but not in the hands of those who believe it is about games, computers, deep mind and deeper mathematical techniques. The beauty of mathematics is that it is fully democratized. However, to add value to an organization, it has to be combined with many other attributes.

Friday, July 26, 2019

Closed loop

As Asian countries send back western garbage to the country of origin, it may be time to think about recycling differently. From water to plastic, humans have been conditioned to think that the cost of consumption is close to zero. But, unfortunately it is not. In the absence of zero cost energy, perhaps hundreds of years away, consumption of anything is costly for the individual and society. It is time we imposed closed loop as a constraint on all consumption.

Closed loop is a necessary condition for survival for humanity. From inception, humans knew this as there is sufficient evidence that ancient societies implemented closed loop systems, perhaps because of lack of availability of critical ingredients. As they grew richer, lazier and greedy, their actions have started to substantially push costs outside their life horizons. This free rider problem coupled with ignorance has reached a critical state. Unless we figure out how to recycle every unit of consumption, garbage is going to pile up in the neighborhoods and oceans. It is time the best minds in the country focused on this problem - AI and space travel could wait, but garbage cannot. Every county in the country may need to be forced to come up with solutions to implement closed loop within their localities. Every time a drop of water is used and a bottle is opened, the county has to prove a closed loop condition, i.e., the consumed unit is  recycled within a closed loop. Technologies exist but not the knowledge and the will to implement them.

Perhaps returning garbage may provide us the impetus to solve this problem for ever.



Thursday, July 25, 2019

Shape of the brain

Research from the University of Chicago (1) that demonstrates brain scans that show distinct differences between criminals who have committed homicide and those who have not, is interesting. If generalizable, such an observation could provide early indications for behavioral health issues. Mental health, often overlooked and treated separately from physical health harbors significant risks for individuals and society. It is a powerful early diagnostic and until healthcare policies treat an individual holistically and not attempt to segment interventions in convenient and legacy packages, we will continue to slip down the quality to cost ratio.

The complex organ humans carry on their shoulders that is speculated to be a quantum computer, is likely the origin of most disease states. As life sciences companies try to tease out meager therapeutic indexes in an overwhelming ocean of placebo effect, it is clear that no therapy can be effective if it is unable to influence the brain. This energy hungry and inefficient mechanism is finely balanced on a knife edge. It is actually surprising we find billions of properly working samples around the world, given the complex design and high maintenance requirements. Prescriptive medicine has laid down strong demarcations between what can be physically intervened with and those less so. Attempts at "curing," a malfunctioning brain using the same principles of crude chemical interventions have led many astray.

If we can understand the brain electronically and magnetically, it could open up a new avenue for diagnosis and treatment. However, it will require educational institutions, providers and manufacturers to work together. But then, that remains to be a dream.

(1) https://news.uchicago.edu/story/scientists-studied-brains-more-800-prisoners-heres-what-they-found

Sunday, July 21, 2019

HOLiCOW (1)

As Hubble turns in his grave, the younger cosmologists and astrophysicists have been making measurements with the latest toys they are afforded. There are doctoral theses to be published and defended, papers to be churned out, tenures and the ever eluding Nobel prizes to be secured. So, clearly better instruments, measurements and debates will help. And none of the budding brains will ever admit that the theory they are banking on is utterly wrong.

The universe was shrinking to a big crunch a few short decades ago and now, it appears to be "certainly," expanding with ensuing debates on the rate of expansion. They send telescopes and satellites that can measure all sorts of data into space, and perhaps that is the problem. When one gets too much data, it becomes difficult to see noise and given the inherent bias to find something and to confirm and conform, most of the latest crop of scientists take the easiest path possible. The fact that they are finding conflicting data is no surprise but what they have to understand is that no new theory will emanate from data. It takes imagination and pure talent, not the type that is built up in the quadrangle by systematic education. In fact, that will assure that none of these geniuses will ever find the truth.

Physics, based on century old theories and an abundance of inexplicable observations by experimentalists may need a reset. Unfortunately, the contemporary educational systems will not allow that.

(1) https://www.sciencemag.org/news/2019/07/debate-intensifies-over-speed-expanding-universe



Sunday, July 14, 2019

Cricket wins again

As England lifts the 2019 world cup in cricket for the first time, the fans of the sport around the world, approaching 3 billion, immerse in joy and sadness, depending on which side one was on. But that does not matter as the game wins hands down for those at the home of cricket and those around the world glued to televisions and short wave radios witnessed an absolute gem of a game. The game has lifted the psyche of the weak and the weary, the quiet and humble, the mathematicians and ballet watchers, the ones escaping boredom and those who do not have time to be bored, the historians and those who recently arrived, the islands that suffered from terrorism and those who attempt to pull themselves up from decade long wars, the ones who are reaching for each other's throats and those who remain far from each other, those who are exiting and those who do not want to, the gun less and those less so, men and women, and everybody else.

Such is the power of twenty yards of dirt and a white ball that travels at 100 miles/hour that it brought tears, laughter and pin drop silences to an oval shaped real estate at the heart of London. There two countries clashed, neither of which has won the cup yet, for pride and enjoyment. Those from the South, trying to wash away the scars of recent horrors in their beautiful island, wanted it bad as it had escaped them just 4 years ago. The larger host, with an inexplicable absence on top of the table in a dozen attempts, had to arrive. The gravity of the occasion was palpable. As the drama unfolded in unpredictable English weather, as the umpires revealed themselves to be purely human, as luck began to make a significant influence on eventual outcome, hearts were broken and unbounded joy erupted. As the winner and the loser thanked each other, it was clear we have a sport for the future, where winning is not everything and losing is just the start.

As we await a world that is civilized and humble - a world that understands there is more to life than ego, power and showmanship, we do have a sport that shows signs of how to reach it.

Saturday, July 13, 2019

The inverted U of closed biological systems

Many have been worried about the existential threats to humanity – asteroids, climate change and ignorant leaders. However, there could be a more systemic and possibly unavoidable threat as long as Earth and its inhabitants continue to be a closed system. It is difficult to hypothesize based on a single observation of the emergence of life in the universe but the theory of evolution appears unassailable. Just as the accepted big bang theory in cosmology that seems to explain a lot after the point of origination but not the event itself, evolution seems to explain most observed biological specimens and their interactions but not the origination of life itself.

The accepted principle of evolution – the survival of the fittest – seems to be fine-tuned to optimize in short horizons. This is understandable as the survival of a species depends on adaptation to the present conditions and not a set of forecasted and uncertain conditions in the future. Thus, evolution creates fitter biological entities in the short run. It has led to more complex entities over time, the reasons of which are less clear. It is possible that the process is driven by underlying laws of physics, that has an arrow of complexity in biological entities aligned with that of time. This may be a side effect of unilaterally increasing entropy in the universe.

Thus, biological entities on Earth have been increasing in complexity. However, more complex entities also seem more fragile – with single cell organisms showing highest level of robustness and the latest iteration in complexity, the human, showing the lowest. In the past, we see several cycles where catastrophic events eliminated complex entities and returned the system to a much simpler state, albeit with differing initial conditions. This implies two things – closed biological systems in steady state will create more and more complex specimens and unavoidable and random catastrophic events will wipe out the more complex members and return the system to a simpler state at regular intervals.

Closed biological systems, thus, may have an inverted U curve on complexity and that may repeat over time. Humans, arguably the most complex in the contemporary system, could be primed for extinction in any large event. What have not been observed in the past are self-destructive internally generated events that humans appear to be very capable of initiating. So, the probability of cataclysmic events now is a lot higher than the past.  The laws of physics may be driving complexity of entities in a closed system higher over time, the laws of nature may be building complex but more fragile systems over time and in the presence of a catastrophic event, internal or external, the most complex entities are removed returning the system to a simpler state with different initial conditions.

The Earth may be due for a reset. This could also provide an alternative explanation for the Fermi Paradox.

Wednesday, July 3, 2019

Weird stars and aliens

As astrophysicists find weird observations, the fallback position for an explanation appears to be the presence of aliens (1). It is about time, as green women have so far failed to show themselves in spite of the space agency's assertion that they would be found before 2020. So, it is important to speculate about the presence of aliens on anything one could not explain by status quo theories and expectations. What the "smart humans," appear to be missing is that civilizations that could run circles around nearby hydrogen furnace would be so technologically advanced that either they make contact or hide from the low life.

The inexplicable dimming and brightening of HD 139139 (1) do not appear to fit with what humans have found and know. Hence they argue there could be aliens there as the first instinctual reaction. Ego and ignorance have the same fingerprint for those harboring the former will never let go of what they thought they knew and those suffering from the latter will never ask questions. Such is the state of science that new entrants to the field shall be brainwashed to believe in what is "known," and trained to fit observations to the accepted theory. It will be anathematic to challenge the status quo as doctoral defenses, tenures, and even careers will be lost in the courtyard, surrounded by ivy walls.

The two axes, the desire to prove life exists elsewhere as proclaimed by those "in the know," and the requirement to fit data to theory (and not dream of a different theory), hold humanity back. As science deteriorates to a level of comedy and speculation, there are real costs to humanity. Sending a billion $ "dragonfly," to the famous moon of Saturn is not an accomplishment, just a revelation that technology has not advanced enough to delegate the source as lifeless. Shooting robots at planets and satellites is not really science, it is engineering gone bad.

It is an intellectual zoo. Most do not want to let go of a theory that is over a century old. And, they are surrounded by engineers, standing ready to prove anything that is hypothesized - aliens and all.

(1) https://www.scientificamerican.com/article/astronomers-dont-know-what-to-make-of-this-incredibly-bizarre-star/