Gill Eapen's Blog, page 66
July 13, 2012
Look ma, no hands
Recent research (1) demonstrates how humans can interact with computers by just using their eyes. It is interesting to ask why computers were designed at all with keyboard, mouse, monitor and such archaic constructs. Why do humans prefer to use their fingers instead of eyes and gestures to interact with machines? The answer simply is that bad industrial designs have conditioned people to expect less.
Carpal tunnel syndrome – the same issue faced by hunters and farmers from thousands of years ago is now prevalent in modern humans as they glibly type away on their keyboards because they were never given any other option. Intellectual giants of our times have designed personal computers and phones with touch screens – but they never asked why? They were following a chain of thought that prescribed the fundamental components of computing. Nobody had sufficient guts and capital to deviate from the status-quo. Bearers of fruits, goblins and micros may one day realize that they have taken humanity in the wrong direction with a dead end.
It is time to fundamentally challenge the basic designs of computing and human-machine interactions. With any luck, we can take the engineers through remedial training.
(1) Controlling your computer with your eyes. Published: Thursday, July 12, 2012 - 23:02 in Mathematics & Economics
July 10, 2012
Failing networks
Advancing analytical techniques are beginning to focus on the prediction of the behavior of systems as opposed to individual components. Although this has been routine in engineering design for nearly half a century, such ideas have not migrated fast enough to business. Lack of sophistication in business management coupled with an overweighted focus on tactics have resulted in dramatic failures of large enterprises and the remaining ones are on their way to the same place. Considering the enterprise as a network of ideas, skills, people and machines as opposed to accounting statements will be a step forward to building the next generation companies (2).
A new crop of leaders with a good understanding of network effects in complex systems are needed to salvage what is left of the large companies.
(1) Cyberwarfare, conservation and disease prevention could benefit from MU researcher's network model. Published: Tuesday, July 10, 2012 - 21:04 in Mathematics & Economics. Source: University of Missouri-Columbia
(2) Flexibility : Flexible Companies for the Uncertain World.http://www.amazon.com/Flexibility-Flexible-Companies-Uncertain-World/dp/1439816328/ref=sr_1_2?ie=UTF8&qid=1341976171&sr=8-2&keywords=gill+eapen
July 8, 2012
Lie-Bore
For over a decade, it has been clear that the financial system around the world is sick. Although productivity in the macro-economy continued to improve during this time, it has not affected prices of financial assets – paper that is being traded by the crooks in glass towers in major cities. The primary reason is the loss of confidence of the public in both the regulatory regime in force as well as the market makers. The rule makers have aided this behavior by making two sets of rules – one for the big crooks and another for everybody else.
It is time this stopped. Both the regulators and the perpetrators have to take full responsibility for what has happened and what continues to happen. The “financial invention,” of this decade – the idea that some institutions are “too big to fail,” is at the heart of this stalemate. If the regulatory regime differentiates between market participants based on their size, most will agree that it is fraught with the problem of moral hazard. The cost of this implicit bailout guarantee is enormous and it has kept a lid on the value of financial assets around the world – making capital expensive for innovators. This in turn has slowed innovation in real markets. With capital drying up – the center of gravity of R&D shifted back to the behemoths with cheaper money but they have a dismal record in innovation. This feedback between financial and real markets may amplify further with recent events and this has disastrous consequences for the economy.
Regulatory incompetence and the lack of consistent application of market principles are destroying the dreams of an entire generation. Tackling this has to be the highest policy priority.
July 4, 2012
Anticipated discovery
The well anticipated outcome at the CERN (1) has arrived – proof that if one keeps looking for the needle in the haystack, one is highly likely to find it – at the very least, some probability can be assigned that the needle in fact exists somewhere in the haystack. This is good news for many, especially those engaged in the hunt. For the rest, it provides relief with the understanding that the universe is orderly and the “standard model,” is “true.” Confirming what was expected is indeed comforting to all. The fact that the standard model is unable to explain most of the universe should not be relevant now as we celebrate the discovery of the last of the dirty dozen that proves the status-quo theory.
“Nobel prizes all around,” said one scientist and yet another cautioned against any caution that may exist – “In my view, if it looks like the Higgs, smells like the Higgs and is exactly what we expected from the Higgs, then it's the Higgs," he said. This is a profound statement and it sheds light on the state of contemporary research. This is a fertile area – as high energy Physics pushes closer to the moment of creation, the speculated fields and particles become increasingly exotic. The more exotic the particles are, the more likely they will be found – as brute force energy and data are highly likely to reveal “what is expected.” The momentum is on the side of those who are making bigger bins and machines and this is a game that cannot be won by waning imagination. 20 Million Gigs of data per year can almost prove anything and is much more powerful than any new ideas. Triple the data rate, and you will be on your way to the “ground-breaking” discovery of “a more exotic version of Higgs.”
Let’s celebrate the discovery today – perhaps tomorrow somebody will ask if it means anything.
(1)“It’s a boson!”: Higgs quest finds what looks like the “God particle”. By Reuters Staff. July 4, 2012
July 2, 2012
Less Pink, more Floyd
After a long drought of 20 years, I watched Roger Waters in concert again, last week. It was satisfying. I gained two major insights from the trip in addition to the music. First, good music does transcend generations – as the age of the audience was bimodal and many families attended together. And, second, we are close to ending an era in which music had meaning.
Pink Floyd defined music for an entire generation – an experience that is unlikely to be replicated ever again. It was a true innovation with significant leaps in many different dimensions – technology, thematic rock & roll and story telling. It defied the status-quo and it attempted to lift mediocre men and women above what was visible and audible. It described the futility of war and the boredom of peace. It exposed the predictability of the rich and the powerful and the inevitability of pain and death. It went further in thought and imagination than anything before it and encapsulated the entire experience in clear and consistent melody. It was music, for sure.
Recent research (1) describes the process of natural selection in music. Natural selection, however, does not guarantee optimal outcomes nor does it promote innovation. Natural selection, by definition, is a process that promotes the status-quo and kills anything that does not fit the present. In essence, it is a process that drives toward average and mediocre outcomes. In a world fueled by natural selection, we are unlikely to see another Pink Floyd, again.
(1) On the origin of music by means of natural selection. Published: Tuesday, June 19, 2012 - 23:31 in Mathematics & Economics
Listen to Dr Bob MacCallum explaining the evolution of evolution of musical loops created by DarwinTunes, accompanied by a selection of loops from the website (mp3 download): https://icseclzt.cc.ic.ac.uk/pickup.php?claimID=84iHGVvekb442b2m&claimPasscode=tPtY3t77PuwENq2S&emailAddr=s.levey%40imperial.ac.uk
Source: Imperial College London
June 25, 2012
Animal instincts
A recent study (1) demonstrates that the location of a crime can be predicted using simplistic models. In this case, the location of gang crimes can be pinpointed nearly 90% of the time – using information derived from the behavior patterns of territorial animals. Humans, thus, have not been able to escape animal instincts in spite of their immense pride in their own special intelligence and scientific developments.
This is a sobering finding at multiple levels. First, the cultural mask that envelope the modern humans is as thin as a veil, for their basic instincts remained intact for over 50,000 years. We may have learned to fake and hide, talk philosophy and toil in the labs – but none of these provide any level of escape from the tribal behaviors that have been part of homo-sapiens from inception. The fact that the probability of a crime can be predicted at the boundaries of gang territories with near certainty means that the macro struggles between countries and religions are replicated at every level underneath – to gangs, communities and homes. Humans, thus, remain and behave like animals, with little change in the fundamental architecture.
Second, the finding implies that humans are highly predictable regardless of where they are and what they do. The rules that make a human are few and the pretense of sophistication is fleeting. Humans, thus, have exposed themselves to an external observer, as a system governed by a few deterministic rules – something that can be run on a simple programmable calculator. It will not take an extra-terrestrial entity much time or energy to figure out how to encapsulate and control humanity.
And finally, the fact that time has not had much impact on humans and their thought processes imply that they are essentially stuck without the possibility of development. They have been around for a long time and even though they have postulated over a 100 “fundamental particles,” stretched life by many days for some, made aluminum tubes that travel faster than sound and erected structures that penetrate the clouds – none of these mean anything at a fundamental level. They will float and coast for many centuries and they will fight like animals till a meteor hit will put them out of their misery.
(1) Remapping gang turf: Math model shows crimes cluster on borders between rivals. Published: Monday, June 25, 2012 - 09:07 in Mathematics & Economics.Source: University of California - Los Angeles
June 23, 2012
Interacting intelligence
Recent work (1) by researchers at MIT and ABB show a more productive path for machine intelligence. Learning machines have been the holy grail of computing but applications have been trivial, thus far. Systems have been mostly limited to mechanical automatons fancifully called, robots. Machine learning has been a broad area – ranging from statistical models to artificial vision. But applications have been heavily dominated by rules – humans unable to shed the limitations of their own wiring. A more practical interaction between humans and machines in which machines are allowed to learn from the rather fickle humans may be better than prescriptive teaching by humans.
Prescriptive languages – that dominate today’s computing – are ill-equipped to implement learning robustly. Early attempts at descriptive languages such as PROLOG were promising but could not get adoption at scale. Every time something new appeared, programmers quickly went back to the environment that is rules based. One of the reasons for this is that engineering education is rules based and this is replicated up and down the whole computing stack.
To get out of this catch 22, two important things need to happen. First, a new generation of computing technologies are needed that fundamentally break away from prescriptive programming. Scientists and programmers have to get comfortable with the notion of a descriptive architecture that does not allow outcomes that are complete and deterministic. And, second, the education systems need to change – with computing making a break from engineering and moving toward the studies of more complex systems such as in Biology and Physics.
Without a major conceptual change in the computing architecture, machine learning will remain a hobby for the conceivable future.
Robotic assistants may adapt to humans in the factory. Published: Wednesday, June 13, 2012 - 11:35 in Mathematics & Economics Source: Massachusetts Institute of Technology
June 15, 2012
Tie the knot
A recent study (1) has found no evidence of cosmic textures in WMAP, casting doubts over contemporary theories. Before the ink dried, scientists have pinned hopes on “much better data” yet to arrive from the Plank satellite. From a purely process perspective, what is the value of a study that contradicts with “established” theory? In most of today’s science, it appears to be zero. If a study does not confirm the accepted notion, just throw out the study and continue looking for “better data.”
This “scientific process,” has now taken over every discipline. We are either in or approaching the dark ages of modern science with the complete absence of the willingness to pursue knowledge. Confirmation and conformation biases are well known, but today’s science has taken it one step further. It categorically denies validity to anything that does not confirm (or conform). This has to stop.
An experiment or a study is not an option. If it reveals a negative result, it has to be as valid as the expected positive result. Intellectual honesty is as important as anything else.
(1) No evidence for 'knots' in space. Published: Wednesday, June 13, 2012 - 11:36 in Astronomy & Space. Source: University College London
June 11, 2012
Logic errors
Contributed by Mr. Ganesh Prasad
June 9, 2012
Skeletons in the closet
Research from Northwestern University’s McCormick School of Engineering and Applied Science at Northwestern Institute on Complex Systems (NICO), demonstrates that complex networks in a variety of domains – biological, technological and social – share the same common underlying skeletal backbone. This fascinating finding has many different applications including understanding disease progression, information dissipation and network optimization in engineering, transportation and electronics. This is further proof that we are entering a regime in which true innovation happens across domains and not within it. Most of the effort currently spent in specialization in medicine, engineering, science and technology are unlikely to result in innovation.
Networks are fundamental – they span every function and every system. From the days of Poincare, over a century ago, it has been postulated that complex outcomes can be achieved by the repetition of standard constructs. The fact that complex networks emanate from the same underlying skeletal systems further reinforce the idea that the universe and everything in it evolve from a few standard templates. Scientific research has been focused on understanding the outcomes and hypothesizing complex theories that accommodate many different observations. In Physics, this has led to the particle zoo – an attempt to explain phenomena by addition rather than by simplification. In Medicine, this has taken the form of treating diseases by inhibiting what may be causing the end outcomes. In Economics, this has led to empiricism in an effort to explain complex behavior and in Engineering, this shows up as incrementalism – an attempt to build on top of the past. Northwestern finding hints that all of these are misguided. What is more important is to understand the templates underlying complex phenomena and not the phenomena itself.
Another important aspect of the finding is that network participants have high consensus on the importance of the links in the system. This also has implications in many different areas including societal and organizational designs. The fact that natural networks exhibit the consensus property universally means that any system designed artificially without such considerations is bound to fail. Highly prescriptive hierarchies, segmentation and segregation, information seclusion and hoarding and other such properties seen commonly in contemporary societies and organizations mean that these systems are not in equilibrium and they will likely break down.
Innovation is a horizontal phenomena. Systems that do not share common features with natural networks are unstable.
(1) Skeleton key, Published: Saturday, June 2, 2012 - 21:32 in Mathematics & Economics. Source: Northwestern University
