Howard Raiffa, Harvard Professor, decision analysis pioneer, dies at 92.

From the Harvard Gazette:

Howard Raiffa, the Frank P. Ramsey Professor Emeritus of Managerial Economics, died July 8 at his home in Arizona following a long battle with Parkinson’s disease.  Raiffa joined the Harvard faculty in 1957. With a diverse group of Harvard stars that included Richard Neustadt, Tom Schelling, Fred Mosteller, and Francis Bator, Raiffa would form the core of what would be the modern Kennedy School (HKS) in 1969, and played a central role in the School for decades as a teacher, scholar, and mentor. Together with colleague Robert Schlaifer, Raiffa wrote the definitive book developing decision analysis, “Applied Statistical Decision Theory,” in 1961. He also wrote a textbook for students like those at HKS, and a simpler, popular book on the subject.

“Along with a handful of other brilliant and dedicated people, Howard figured out what a school of public policy and administration should be in the latter decades of the 20th century, and then he and they created that school,” said Raiffa’s longtime friend and colleague Richard Zeckhauser, Frank Plumpton Ramsey Professor of Political Economy.

“Despite his great accomplishments as a teacher and scholar, those who knew Howard well treasured him for the generosity of his spirit, his great warmth, and his desire to always be helpful, whether fostering cooperation among nations, choosing where to locate Mexico City’s airport, or designing a curriculum for teaching analytic methods.”

This combination of work marks Raiffa as a model for the Kennedy School: His scholarly analysis advanced experts’ understanding of many important questions, and he also knew how important and valuable it was for him to speak to the broader world.  In particular, he recognized that the methods he had pioneered and mastered could be helpful to people with much less sophistication, and he reached out to help them.

“Howard was a giant in the history of the Kennedy School and a towering figure in the fields of decision analysis, negotiation analysis, and game theory,” said HKS Dean Douglas Elmendorf. “All of us who are associated with the Kennedy School are greatly in his debt.”

More on Howard Raiffa:

• New York Times Article 
• Books on Amazon
• A conversation with Howard Raiffa in Statistical Science  (PDF)
• Introductory lecture on decision under uncertainty  – (PDF)
• Video (below)

Are we living in a simulated world?

That’s right!, there is a debate going on in some scientific and not so scientific circles whether we are living in a simulation, just like Super Mario.  Warning: this post may introduce a bunch of ideas you may have never heard about…

Where did this idea come from?

In 2001, Nick Bostrom, a Swedish philosopher based at the University of Oxford, came up with a paper entitled  “Are You Living In a Computer Simulation?“. He actually published it in 2003.  The paper is in part inspired by ideas that developed in the science fiction, futurology and philosophy world, including post-humanism, “big world”,  and terraforming. Let me explain quickly what that is about… some of this stuff is a bit “out there…”.

• Post-humanism is about what follows the human race, or what evolves from the human race. You could think of intelligent cyborgs, or robots that would take over from us.
• Terraforming is about conquering and establishing human life on other planets.
• The Big World is a universe with macroscopic superposition, where entire worlds like us are superposed onto one another.

Bostrom proposed that there is a significant chance that any of the three following scenarios are true: 1) the human race will go extinct before becoming post-human, 2) A post-human society will not run simulations of its evolution, or 3) we are currently living in a simulation.  It must be one of these three, says Bostrom, and he shows that the probability of each choice is significant, i.e. not zero.  Bostrom did not start this discussion. Decades before Bostrom’s mathematical argument, folks like Jacques Vallee, John Keel, Stephen Wolfram, Rudy Rucker, and Hans Moravec explored this notion.

What do they mean by “simulation”?

Simulation means that the rules under which we exist and by which we live are controlled by a machine. The simulation is a “computer program” that makes things in our universe happen.  This implies, of course, that there is another universe, or reality, in which this controlling program and its computer exist.  I will explain below why I put “computer program” in quotes.  Of course this is a total supposition.  The simulation could have set the basic rules of life and evolution, from which the World has evolved.  Or the simulation could actually control every step of our existence and the movement of every particle known in the creation.  Is the simulation program perfect? Some writers have argued that the simulators (the ones controlling the simulation) may have the ability to erase any program errors from our memory, or that we could not even perceive these program errors if they exist…

Why is this idea even debated these days?

This topic is debated in pretty serious circles, like the Isaac Asimov Memorial Debate, recently held at New York’s Hayden Planetarium.

We live in an increasingly digital world, where computers are getting faster, bigger and cheaper all the time.  Video games computer-generated movies (simulations) are becoming more life-like.  Computer-driven machines (post-human robots?) are becoming more human-like.   Movies like the Matrix and the Truman Show have popularized the notion of outside worlds, or worlds within worlds.

But there are other interesting developments in the world of physics that make this idea intriguing.  Over the centuries, we have tried to explain the world around us through science.  We have built tools to observe our world further and further, within and out.  We discovered and proved the existence of atoms and electrons in 1897. In 1932, we found about neutrons.  In 1962, physicists started to talk about quarks.  We are now looking for evidence of the famous Higgs boson. It takes very sophisticated and expensive tools such as linear accelerators and the Large Hadron Collider (LHC), the world’s largest and most powerful particle collider, to study what our universe is made of.   At the same time, we are learning about dark matter and dark energy that may explain how the universe is expanding.

Evidence increasingly points to the possibility that our world may be made of waves and bits. Matter, the one we see around us, may also ultimately be bits, wave and information. One of the first to advance such a theory was Edward Fredkin, a professor at MIT, then Boston University and Carnegie Mellon.  Back in the sixties, he came up with the theory of digital physics, postulating that information is more fundamental than matter and energy. He stated that atoms, electrons, and quarks consist ultimately of bits–binary units of information.   So this is not a completely new story.

Mathematicians and philosophers are exploring such theories because our scientific tools are not capable of looking smaller or further for now.  So, if we are starting to believe in a digital world made of ones and zeros, it is no longer a giant leap to think of it as a giant computer, or being driven by a computer program.  Take a look at Kevin Kelly’s 2002 article, “God is the Machine.”

Simulation: yes or no?

So, are we living in a simulation? There are a lot of discussions and arguments out there about whether we are in a simulation or not.  Some people argue that it would take too much energy to run the simulation of our world and perhaps other simulations running at the same time.  Imagine walking on a beach…every grain of sand would have to be part of this simulation, moving in and out of the ocean.  Think about every encounter and every discussion with another human being being pre-scripted..

My personal view is that this idea that we live in a simulation is improbable. One main reason is that all of these ideas and concepts are the product of our language and our ability to reason using language.  Mathematics is another form of language and reasoning. Language and conceptual reasoning are the product of our limited human abilities.  The concept of “simulation” is something that we can grasp. But what about a million other concepts that we cannot grasp or formulate.  What about a million other forms of intelligence out there.  Zeroes and ones, the alleged basis of our universe, are mere human inventions.  It is possible that we live one of several forms of reality, but the “simulation” idea is simplistic to me.  There is likely something out there so different that we cannot even express with our limited intellect or language…

Along the same line (or on the other hand)…  describing the universe as pure information may also be a simplification.  Life as we know it runs on some fundamental rules: the first one is reproduction, the concept that life has a built-in reproduction mechanism,  the second concept is that of healing. Most living organisms are able to sense when they are hurt or attacked and are able to react with a plan to heal themselves. The third rule of life is that organisms evolve and adapt to their environment in order to survive.  Something, somehow, has come up with these rules…

Organizational complexity is taking a toll on business profits

Last July and August 2015, the Economist magazine conducted a survey of 331 executives. The survey focused  on the impacts of organizational complexity and the efforts companies have taken to reduce it.

Complexity in Large Organizations

Large organizations are viewed as dynamic networks of interactions, and their relationships go well beyond aggregations of individual static entities. For instance, these organizations may operate in many countries and manage a vast number of people and brands.  But the perception of complexity can also stem from a lack of employee role clarity or poor processes.

Unwieldy complexity often results from business expansions or bureaucracies that unnecessarily complicate a company’s operating model, leading to sluggish growth, higher costs and poor returns.  The Economist survey found that over half (55%) of businesses suffer from organisational complexity which is perceived to take a toll on profits.   Most affected by complexity is general management, followed by employee relations and customer service.  Over one third of executives said that they are spending up to 25% of their time managing that complexity, and almost one in five is spending up to 50% of their workday dealing with complexity.

Two kinds of complexities

Experts believe that complexity in the organization comes from two sources:  The complexity arising from the outside World, its dynamic and unpredictability, and a company’s internal complexity caused by poor processes, confusing role definitions, or unclear accountabilities.  The most damaging kind of complexity come from within.  In a survey of 58 companies, Mocker and Jeanne Ross, Director and Principal Research Scientist at CISR, found that those companies that are able to create value from product complexity while maintaining simple processes were outperforming others.

According to Mocker, companies can then boost value and organizational effectiveness through a combination of tree sets of actions:

• Break the separation between those dealing with product complexity and those dealing with process complexity,
• Design processes and systems to cushion internal impacts of complexity by eliminating silos, and
• Offer customers ways to deal with increased complexity by offering personalized choic es for instance.

John Holland, ‘Father of Complexity,’ Dies at 86

Reposted from https://reason.com/blog/2015/08/31/john-holland-father-of-complexity-dies-a

Pioneer in complex adaptive systems passed away earlier this month

by Ed Krayewski|Aug. 31, 2015 7:15 am

University of Michigan

Earlier this month, the computer scientist and psychologist John Holland passed away. Holland is known as the “father of complexity” for being the pioneer in the study of complexity and complex adaptive systems. He studied under Arthur Burks, who worked with the visionary computer scientist John Von Neumann, whom he helped design the first electronic digital computer, and who came up with the idea of the Von Neumann machine, a self-reproducing computer program, before DNA was discovered.

Scott Page writes at the Washington Post:

Holland was fascinated with von Neumann’s “creatures” and began wrestling with the challenge and potential of algorithmic analogs of natural processes. He was not alone. Many pioneers in computer science saw computers as a metaphor for the brain.

Holland did as well, but his original contribution was to view computation through a far more general lens. He saw collections of computational entities as potentially representing any complex adaptive system, whether that might be the brain, ant colonies, or cities.

His pursuit became a field. In brief, “complex adaptive systems” refer to diverse interacting adaptive parts that are capable of emergent collective behavior. The term emergence, to quote Nobel-winning physicist Phil Anderson’s influential article, captures those instances where “more is different.” Computation in the brain is an example of emergence. So is the collective behavior of an ant colony. To borrow physicist John Wheeler’s turn of phrase, Holland was interested in understanding “it from bit.'”

Read the rest of Page’s write-up at the Post.

Readers interested in introducing themselves to Holland should read Signals and Boundaries: Building Blocks for Complex Adaptive Systems, which applies the ideas of complexity to biology, markets, and even governments, and vice versa.

Other Resources

Slideshare

• John Henry Holland by Lucia Artigas
• Complex Adaptive Systems Theory
• Complex Adaptive Systems and Communities

Articles

• Santa Fe Institute:  Complexity science giant John Holland passes away at 86
• PBS: Emergence: Expert Q&A
• New York Times: John Henry Holland, Who Computerized Evolution, Dies at 86
• University of Michigan:  Obituary: John Henry Holland
• Wikipedia: John Henry Holland
• Britannica: John Henry Holland

Systems Engineering Standards to Help Manage Complexity

Back in 2009, an IEEE Spectrum article noted that the avionics system of the U.S. F-22 Raptor jet fighter launched in 1995 consists of about 1.7 million lines of software code. Today, premium cars like the Mercedes-Benz S-class depend on millions of lines of code running up to 100 networked electronic control units (ECU) throughout the body of the car, controlling and monitoring everything from the powertrain to the airbag system.  As vehicles grow in complexity, coordination between carmakers, suppliers and other partners faces increasing challenges.

Safety recalls for automobiles sold in the U.S. are soaring. Part of the reason is the growing use of common parts across multiple model lines, and even by multiple automakers. Automakers have to connect with international suppliers, joint ventures and automaker partners operating different systems.  One defective part may now affect multiple brands.

Siegmar Haasis, CIO for Research and Development at Mercedes-Benz Cars, believes that standards can help manage the increasing complexity inside the company as well as outside the company.  One way to manage the increasing automotive design complexity is the use of ReqIF (Requirements Interchange Format) to facilitate the exchange of specifications for parts and systems between partners. Daimler is also testing systems engineering standards like OSLC (Open Services for Lifecycle Collaboration), FMI (Functional Mock-up Interface) , XMI (XML Metadata Interchange) and SysML (Systems Modeling Language).

The wide use of standards can be transformative. A good example is the ISO shipping container, now accounting for 60% of ocean-going freight.  Standard containers have enabled intermodal transportation, resulting in the development of modern supply chains and global trade. Standards make complexity more manageable, enable parallel work and accommodate uncertainty about the future.

But standards have drawbacks too.  Standard adoption can damage competition and innovation.  A standard is partly designed to limit or “stop” innovation (or dis-order) for a while in order to rationalize systems design, production, or operation.  Ideally, standards should enable the inclusion of new knowledge over time, but that, by definition, is hard to do.  Standards and their adoption affect company strategy and market share over time.   Like most things connected today, the impact of standards on systems engineering can be… complex!

Beauty as a Concept

Most of us would agree that watching the sun set over the horizon is a thing of beauty. In an instant, we are leaving our complex world behind and feel at peace for a moment or two…

Beauty as an Emotional Experience

What is it about a perfect sunset that makes us feel that way?  or why are we disappointed when clouds obscure the sunset’s final moments above the horizon?  Lots of folks have discussed the nature of “beauty”and how it holds our mind’s attention.  Researchers have found that people have similar beauty standards across China, India and the USA.  Back in the 1750’s, the German philosopher Immanuel Kant distinguished between beauty as an intuitive or sensible experience and a rational approach to aesthetics on the other.   In the first instance, beauty is an emotion.  In the second instance, we are trying to use logic to establish whether something- a painting, a song or a piece of clothing – is beautiful. Clearly, watching the sunset is an emotional experience, beyond logic.

Is Logic All We’ve Got?

Science and engineering are purely based on logic.  In fact, logic is the main human language of reasoning. The entire “digital” world of computers is based on logic. Moore and Parker argue that logic is the mode in which our mind reasons best about our reality. But there is another important way to think, and analyze, and feel: conceptual thinking.  I covered conceptual thinking briefly in another post.  Every day, we build, update or modify complex, flexible, abstract representations of reality called “concepts”. For instance, we recognize just about any cat by matching it to our concept of “cat”, even if the cat wears a costume!  Psychologists have tried to understand how we build these abstract models. Jean Piaget, the famous Swiss psychologist developed a theory about the nature and development of human intelligence which includes observations about how children develop concepts.  In general, however, these analyses fall short of explaining the mathematical and physical structure of concepts, how they are learned and how they work.

Beauty as a Concept

Conceptual thinking may not be as definitive or as accurate as logic (although logic based on bad assumptions will yield bad answers!) but it is much faster. We can figure out with near certainty that something is a cat much faster with conceptual thinking than with logic: Does it have fur? Does it have whiskers? Is its size compatible with that of a cat? Is it not a dog? Etc…

When we see the sun setting over the horizon, the scene before us, no matter where it is, reduces itself to a simple, timeless, abstract and moving concept:  beauty.

Three skills for dealing with complexity

Our World has generally become more complex, in part because of the many linkages and relationships between elements that make it a lot harder to figure out how things work and where they are going.  Let’s take the stock market for instance.  One day it is affected by the country of Greece, and the next by Amazon’s quarterly revenue. The price of an airline ticket to a particular destination is another good example. It depends on the time of year, the day of the week, the flight’s time of day, the airline’s particular network structure, the number of stops, how many seats remain in each fare class, what the competition charges, what the market will bear, etc…  Below are three important skills to help us deal with complexity:

Critical Thinking

Critical thinking is the ability to break down a complex system (my definition) and using analysis to understand its inner-workings.  Critical thinking requires research, critique, analysis and evaluation of the system and its sub-components. “Research” means gathering facts, information and opinions. “Critique” is the ability to discern what is important and what is not.  “Analysis” (from the Greek “αναλυειν”, meaning to “untie”) is the ability to understand how things work.  “Evaluate” means forming a quantitative assessment of a system, understanding how it may behave under various circumstances. Critical thinking is about identifying what, in the complex system under study, is important for the particular situation at hand.

Conceptual Thinking

We all know this is a cat!

This is a tough one.  Conceptual thinking is the ability to reduce a complex system into a simple, critical model.  The concept of a cat, for instance, is a representation of knowledge in our brain about what makes a cat, and how a cat is different from a dog, or an elephant. Cats may be all very different, but we recognize them all as cats and not dogs.  We know it when we see it!   Things are a bit more complicated when we talk about complex systems.  Both experience and creative thinking help reduce a complex system into a simple conceptual model. We then intuitively understand how it works.   Take for instance the concept of “supply and demand”.   Conceptual thinking means that we recognize that the price of an airline ticket is highly driven by supply and demand.  Suddenly, we understand better what drives the price of an airline ticket.

Risk Management

Reducing a system to its critical relationships, or building a conceptual model to understand it, are key to decision-making and to management. Of course, it is possible that we do not understand perfectly the complex system we are looking at or that we are not able to predict how it will behave.  In particular, assumptions about what will happen in the future are likely be shattered by unforeseen developments.  That’s where risk management comes in.  Risk management is a systematic process of  addressing and understanding what can go wrong with our assumptions.  What can be done?  Typically we can build flexibility into our decisions, but that will likely cost us something.  For instance, we can buy a refundable air ticket in case we change our mind, but that’s more expensive as we know.

Thinking critically, conceptually and systematically addressing risks are important, lifelong skills that everyone should learn.