Broken Symmetry

Wolfram on "Facebook World"

Very interesting article and graphics here:

I have to say that as I look at the plots above, I’m struck by their similarity to plots for physical processes like chemical reactions. It’s as if all those humans, with all the complexities of their lives, still behave in aggregate a bit like molecules—with certain “reaction rates” to enter into relationships, marry, etc.

Eddies in Memory

Suppose you're allocating computer memory to local variables associated with a host of objects (maybe millions). You need to free up that memory when an object is no longer needed. So you need a way to be able to tell whether the memory still has any local variables still in use pointing to it.

This can be visualized as a directed graph. The set of nodes with only outgoing edges are the variables still in use. The nodes with no incoming edges represent the memory no longer in use.

And it appears that at any given moment, the set of nodes with no incoming edges will always form a cycle -- eddies in the directed graph that represents your memory allocation.

For some reason, I find that aesthetically satisfying.

This Day in History: University Research and the Internet

From the Computer History Museum:

March 16, 1990

Internet Extends Beyond U.S. to Europe

The National Science Foundation announces it will extend its network with a high-speed data link to Europe. Five years earlier, the Internet in its modern form had started to develop rapidly thanks to the formation of the NSFNET, which linked five supercomputer centers in the United States. Later in 1990, Europe contributed to the growth of the Internet when CERN's Tim Berners-Lee developed HTML, the language used for the World Wide Web.

Has any technological development done more to change the world in the same period of time? Why do some folks seem to have forgotten how the Internet got started? There were lots of venture capital funds around in 1990. How many were ponying up to fund a transatlantic link between NSFNET and Europe?

Facebook and Ennui

Rick Bookstaber gives an interesting take on the alternative trajectories of society and Facebook over the next few decades:

I am not going to enter the discussion of how secure Facebook's position is, or how much advertising it can generate. Insread, I am going to discuss the implications of Facebook for our individuality as we increasingly embrace Facebook and similar systems that replace real interaction with the virtual, and the implications of this world for the success of Facebook.

The basic point is that either the world changes and Facebook become marginalized, or Facebook continues [to] thrive and we live in a world of existential alienation. ...

Interestingly, Facebook wins relatively few friends in Japan. I wonder how Bookstaber would interpret this in view of his thesis.

High-Frequency Trading Fact of the Day

From an interesting paper on the flash crash by Easley, Mailoc Lopez de Prado, and O'Hara:

Since 2009 high-frequency trading firms, which represent approximately 2% of the nearly 20,000 trading firms operating in the U.S. markets, have accounted for over 73% of all U.S. equity trading volume.

Via Computational Legal Studies

Fixed Points in the Publishing Industry

From an interview of Lonely Planet's Gus Balbontin on O'Reilly Radar:

The essence of what the print people have known for 500 years is still the same in the digital world: you still have to publish great content. Yes, the tone may change a little. Yes, the experience might be slightly different. But at the end of the day, the people creating the content are humans. The people consuming it are also humans. The same humans that consume print consume digital. It's not different.

...

In terms of the next 10 or 20 years, the essence of what publishers are doing won't change at all. This is what should make publishers feel comfortable about the future. The reason why you're there and the problems that you can solve for customers won't change. That's what you need to keep focusing on.

The Internet Operating System

From Tim O'Reilly:

Might an operating system of the future manage when and how data is collected about individuals, what applications can access it, and how they might use it? Might it not automatically synchronize data between devices and applications? Might it do automatic translation, and automatic format conversion between different media types? Might such an operating system do predictive analytics to collect or locally cache data that it expects an individual user or device to need? Might such an operating system do "garbage collection" not of memory pointers but of outdated data or spam? Might it not perform credit checks before issuing payments and suspend activity for those who violate terms of service?

Tim probably inadvertently sent a lot of traffic my way thanks to link through to a Google search of windows "bag of device drivers" andreessen.

Scalability is ultimately the key determinant to what Internet "operating system" will dominate. Google has the "unfair" advantage at this point. Microsoft also has a big cost disadvantage because of its need to support legacy revenue streams. Apple doesn't seem to have the capabilities now to code scalable server-side apps. iTunes is a nightmare. Cisco's products are complements to the operating system market.

Maybe there is still room for a new competitor to Google. Yahoo has more of an opportunity here than many people realize.

The Topology of Knowledge

From Joho the blog:

In the Age of Links, we still use credentials and rely on authorities. Those are indispensible ways of scaling knowledge, that is, letting us know more than any one of us could authenticate on our own. But, increasingly, credentials and authority work best for vouchsafing commoditized knowledge, the stuff that’s settled and not worth arguing about. At the edges of knowledge — in the analysis and contextualization that journalists nowadays tell us is their real value — we want, need, can have, and expect transparency. Transparency puts within the report itself a way for us to see what assumptions and values may have shaped it, and lets us see the arguments that the report resolved one way and not another. Transparency — the embedded ability to see through the published draft — often gives us more reason to believe a report than the claim of objectivity did.

Consider the following possibility: Maybe transparency was objectivity all along, but it simply took much, much longer for feedback loops in analog social networks to discredit overreaching authorities than it does in digital social networks.

A corollary is that credentials and authority aren't going anywhere. Their relative importance in the offline world just went up precipitously at the same time that it declined precipitously in the online world.

What Network Structure Provides the Smoothest Flow?

From a paper by Zoltan Toroczkai, Balazs Kozma, Kevin E. Bassler, N.W. Hengartner, and G. Korniss:

If the flow received by a node has to be processed, it will happen at a finite rate. For example, a node receiving a packet, has to read off its destination and find out to which neighbor to send it. This is a physical process and takes a non-zero amount of time. Thus, if a node receives too many packets per unit time, they will form a queue, and long queues will generally cause delays in information transmission and thus leads to congestion, or jamming. An important question then arises: Can the topology of the underlying substrate graph influence the level of congestion in the network?

The answer is yes...

See also the wiki summary.

Twitter Updates

Have now been added to Broken Symmetry on the right. If you're already tweeting, be sure to follow me. I'm riemannzeta. If you tell me you read Broken Symmetry, I'll be more likely to reciprocate.

Do Email Traffic Statistics Really Cast Doubt on Small World Theory?

The answer is no. MIT Technology Review reports on work published by researchers at the Hebrew University in Jerusalem:

Goldenberg and Levy say that while it is just as easy to email somebody who lives on the same street as somebody on other other side of the world, it turns out we have a huge preference for sending messages over shorter distances... Their conclusion is that far from reducing the importance of geographical location, electronic communication appears to have increased it, probably because people swap more messages with those they have personal interaction with. If that's true, why have we gone so wrong in thinking the world is getting smaller? One source of confusion, argue Goldenberg and Levy, is the famous six-degrees-of-separation experiments peformed originally by Stanley Milgram with letters and later by Steve Strogatz and Duncan Watts using email. These seem to indicate that a "small world" effect is at work in social networks.

The major problem with this criticism is that it misunderstands the point of the Milgram and Watts-Strogatz work. Milgram, Watts, and Strogatz had little data on the dynamics of the network. They were studying the network structure. This looks like sensationalism to me.

If you asked any of them whether they would be surprised to hear that the frequency-spectrum of communication is spatially clustered, I'm sure they would say no.

Happy Flu

Internet does not exist per se, it is merely the volatile aggregate of information, flowing around. For the past few years - following the emergence of the so-called Web 2.0 - information diffusion has been more and more orchestrated by bloggers instead of mainstream media. This paradigm shift has given us the opportunity to understand how word of mouth works by providing an actual framework in which a resource could be observed while spreading.

That's from happy flu. Be sure to check out the video.

Via the Complexity and Social Networks blog.

CEO of Beth Israel Hospital in Boston Blogs

Check out Running a Hospital.

Our President has a blog.

If you're a leader of any kind, blogging is a very good way to communicate your vision, and to receive feedback in real time from people within your organization.

See also this earlier post on the impact that blogging by leaders might have on innovation.

Whom can you trust?

From Clay Shirky:

To describe the world before or after the spread of print was child’s play; those dates were safely distanced from upheaval. But what was happening in 1500? The hard question Eisenstein’s book asks is “How did we get from the world before the printing press to the world after it? What was the revolution itself like?”

Chaotic, as it turns out. The Bible was translated into local languages; was this an educational boon or the work of the devil? Erotic novels appeared, prompting the same set of questions. Copies of Aristotle and Galen circulated widely, but direct encounter with the relevant texts revealed that the two sources clashed, tarnishing faith in the Ancients. As novelty spread, old institutions seemed exhausted while new ones seemed untrustworthy; as a result, people almost literally didn’t know what to think. If you can’t trust Aristotle, who can you trust?

Our notion of credibility and authority is changing in fundamental ways because of the Internet.

Organic Economics

Since the Arrow-Debreu-McKenzie model was introduced to economics in the 1950s, changes over time have been given less consideration in economic models. See, for example, this statement about changes in market prices by Eugene Fama. The efficient-market hypothesis implies that at any given moment an equilibrium market price exists, and (here's where the flow of time drops out of the picture) that the current price at any given moment is most likely the equilibrium price. Unless you have some information that nobody else in the market has, then you're probably making a mistake by guessing that most people are either underestimating or overestimating the value of a particular good. There are exceptions, of course, but in general economists and policymakers (including judges and lawmakers) think about (or at least did think about prior to this recession) economic equilibrium as static —- a perfect allocation of people and things that the market maintains more or less constantly.

Is there a more realistic way to model an economy? How about as an hierarchical social network? The social network must be hierarchical because smaller markets (i.e., smaller subnetworks) are usually connected to larger markets (i.e., bigger networks) such that there is nesting and overlap among them through the mechanism of cross-elasticity in demand.

Already, our model of the economy is of networks within networks. But the hierarchy goes deeper still. Each individual within an economy has his or her own neural network, which orders decisionmaking and behavior. In principle at least, we could identify subnetworks within an economy by looking to see whether particular individuals share substantially similar neural networks. Similar brain structure and dynamics are probably a physical (and biological) manifestation of consensus in belief and perception about past, present, and future within a subnetwork. Market equilibrium could be identified as the state of the system in which neural networks are matched in structure and dynamics so that the flow of goods through the economy is stable over a period of measurement. Such stability would be manifest in a market price that fluctuates only within the range of transactions costs.

From time to time, however, an individual within a subnetwork might make a new connection in his or her neural network. The new connection would arise from a new structural or dynamical pattern within that individual's neural network. Such a new pattern might be induced directly by exposure to new sensory information or statistically random variations in the individual's neural network, or indirectly from exposure to a new pattern communicated by another outside the subnetwork. In any case, the result would be a modification to the preexisting structure and dynamics of that individual's neural network.

An individual with a new pattern in mind (literally) might then attempt to reproduce the new pattern in the subnetwork of individuals who share (now similar, but not the same) neural network structure and dynamics. Through communication, the new pattern is either recognized, thereby rerouting every neural network in the subnetwork (albeit, through feedback loops, probably with some refinements), not recognized at all because too great a variation from preexisting patterns to be communicated, or rejected as inconsistent with important preexisting accepted patterns. Although a ridiculously difficult problem to solve mathematically, whether a new pattern would be accepted or rejected by the subnetwork could in principle be characterized using control theory to analyze the stability of the system including the new pattern within its preexisting environment. In practice, it's probably necessary to carry out the experiment to know.

The iteration of such a process eventually results in an increasingly sophisticated neural network shared among the subnetwork, with more and more specialized ability to recognize and explain more and more complex information over time. Interconnections among subnetworks, although difficult to achieve, might be expected to produce large returns to the individuals who are capable of matching the patterns in otherwise disconnected subnetworks.

To understand the new insights that such a model would give us in understanding economics, consider how individuals within the network of an economy include both entrepreneurs and customers. A successful entrepreneur, in practice, is one who is able to anticipate what patterns are going to be accepted by the neural networks of customers. Within this organic model of economics, capitalism is a game in which entrepreneurs conduct experiments in order to discover what patterns of behavior (i.e., what neural networks of prospective customers) will adapt by accepting a new pattern offered as an alternative to a preexisting pattern of behavior used to live with time-evolving resource constraints. Correct predictions are rewarded by the mass of individuals that adopts a new pattern of behavior. Incorrect predictions -- or inappropriately timed predictions -- generally disappear without a trace.

Not coincidentially, this game of capitalism provides a social parallel to the flow experience described by Mihalyi Csikszentmiahlyi as optimal to individual psychology. The same basic rules of the game apply both at the level of neural networks and at the level of networks of individuals within an economy. Because of how neural networks constitute economic networks, it should not surprise us to see that similar mathematical structure and dynamics apply to both. Consumer internet marketing professionals probably already know more about how neural networks work than do many neurologists!

Neoclassical theory offers zero insight into any of these dynamics.

Published Patents and Patent Applications Now Available via XML from Data.Gov

See here.

Regardless of how you may feel about Obama administration ideology, the implementation of transparency through data.gov is a huge win for us. I am really excited about the possibilities for more informed interaction with government.

Dynamic Centrality by Bar-Yam and Braha

Available here:

In summary, we have demonstrated that the static topology does not capture the dynamics of social networks. The prominence of nodes (as measured by degree) within the networks fluctuates widely from day to day, and a high degree in the aggregate network does not predict a high degree for individual days. Our conclusions are in sharp contrast to previous complex network research, which emphasizes the importance of aggregate nodal centrality in a static network topology [1– 4,6,10,11,14 –18].

Implications of a dynamic node centrality contrast with existing analyses that consider targeting nodes with the highest degrees to disrupt network communication or transport. Dynamic centrality implies that targeting nodes with the highest degrees at one time only weakly affects the nodes that are highly connected at another time. The approach of targeting high-degree nodes has been suggested, for example, to be an effective disease and computer virus prevention strategy; i.e., identification and “vaccination” of those nodes would inhibit the spread of infection or computer viruses [4,6,19 –22]. Our work implies that, at the very least, a more agile strategy of monitoring and vaccinating nodes based on centrality over time is necessary. Otherwise a treatment based on aggregate connectivity information will miss the impact of a node that otherwise has a low connectivity, becoming highly connected.

Characterizing Individual Communication Patterns by Malmgren, Hofman, Amaral, and Watts

Available here:

The increasing availability of electronic communication data, such as that arising from e-mail exchange, presents social and information scientists with new possibilities for characterizing individual behavior and, by extension, identifying latent structure in human populations. Here, we propose a model of individual e-mail communication that is sufficiently rich to capture meaningful variability across individuals, while remaining simple enough to be interpretable. We show that the model, a cascading non-homogeneous Poisson process, can be formulated as a double-chain hidden Markov model, allowing us to use an efficient inference algorithm to estimate the model parameters from observed data. We then apply this model to two e-mail data sets consisting of 404 and 6,164 users, respectively, that were collected from two universities in different countries and years. We find that the resulting best-estimate parameter distributions for both data sets are surprisingly similar, indicating that at least some features of communication dynamics generalize beyond specific contexts. We also find that variability of individual behavior over time is significantly less than variability across the population, suggesting that individuals can be classified into persistent “types”. We conclude that communication patterns may prove useful as an additional class of attribute data, complementing demographic and network data, for user classification and outlier detection—a point that we illustrate with an interpretable clustering of users based on their inferred model parameters.

Via the Physics arXiv Blog, which has been in sync with my interests very well lately.

Folks, this is how economics looks without the rational hypothesis.

(UPDATE 8/12/9: This is how economics looks with a more granular model of preferences.)

Looking at the two clusters in parameter space, I can't help wondering whether the one is for staff and administrators and the other for students and faculty.

Small Steps Toward Solving Hayek's Knowledge Problem through the PTO

From the consistently interesting Computational Legal Studies blog:

The United States Patent and Trademark Office patent classification scheme organizes 3 million patents into about 160,000 distinct patent classes. This visualization by Katy Börner, Elisha Hardy, Bruce W. Herr II, Todd M. Holloway, & W. Bradford Paley considers the organizational schema used to classify patents at the US Patent Office. Their article Taxonomy Visualization in Support of the Semi-Automatic Validation and Optimization of Organizational Schemas was published in the Journal of Informetrics in 2007.

Internet Traffic Periodicities and Oscillations by Reginald D. Smith

Available here:

User traffic driven periodicities were the first known and most easily recognized. The first discovered and most well-known periodicity is the 24 hour diurnal cycle and its companion cycle of 12 hours. These cycles have been known for decades and reported as early as 1980 and again in 1991 as well as in many subsequent studies. This obviously refers to the 24 hour work-day and its 12-hour second harmonic as well as activity from around the globe. The other major periodicity from human behavior is the week with a period of 7 days and a second harmonic at 3.5 days and barely perceptible third harmonic at 2.3 days.

There is a lot more that could be done with this type of application

http://omegle.com/

Blogging and Scholarship

In general, it seems the academic community is warming up to the idea of blogs as complementary to scholarship. The benefits are obvious enough -- blogs can potentially reach a wider audience and provide feedback on new ideas quicker than the traditional system of private correspondence, conferences, and peer reviewed publication.

There are certain fields -- mathematics, sociology, literature, poetry, philosophy -- for which these benefits seem decisive. Although a scholar in these fields will still want to publish in addition to blogging, once a critical mass of scholars in these fields are publishing their work online, the community within the field will increasingly find the Internet to be a more convenient medium for exchanging ideas. Each scholar has control over their own printing press. Going from expression to wide publication takes zero time rather than months or years.

There are other academic fields, however, for which the benefits seem less decisive. These fields include, for example, experimental physics, engineering, and medicine. These fields, by comparison, are driven more strongly by new empirical results. Empirical studies are time consuming, and take months or years to complete. The time scale of reporting cannot be shorter than the time scale of information gathering.

By the way, neither group of fields is entirely empirical or theoretical. It's just that, on balance, the first group is more theoretical and the second, more empirical. The two groups need each other, and the growth of both fields will be limited by how much information (both data and theory) is shared between them.

In terms of raw numbers, I think the off-line population of the second group is actually much larger than the first group because the second group is generally the group that society calls upon to assist in solving its problems. This is the academic group that has the most direct impact on the rest of society, and its number bears that out. By contrast, the first group (of theorists) probably has a more lasting impact on the way that academia evolves independent of society. The big successes in the first group percolate out of academia through the second group, and eventually change society as a whole. Marx changed the face of the world with his ideas. The Wright Brothers changed the face of transportation. Both theory and practice were necessary to both; but their day to day lives looked quite different, and their influence on society over time reflects those differences.

At the moment, it seems that the first group is going to grow much faster than the second group into the blogosphere. Over the short run, this isn't a big deal. There are still plenty of sources of data available online. The problem is that many of these sources are journalists from newspapers and magazines that are on the verge of bankruptcy. Moreover, very few of these sources of data are put together using the careful methodology that one might find in a typical empirical study by an economist or sociologist, for example. In effect, the Internet is crammed with lots of theorists analyzing a tiny amount of data that is from comparatively unreliable sources. The Internet is a bit like our consciousness in that sense.

What can we do to bring more data online, and keep it organized but accessible? The people who specialize in gathering data now have little incentive to put it onto the Internet when the existing mechanisms for sharing that information are just as good or better, and make it easier to predict whether and how recognition and reward will be distributed tot he gatherers?

People rave about the importance of openness on the Internet. But already I'm beginning to see that the evolution of property rights will be no different on the Internet than it was on any frontier of the past. To encourage people to share valuable goods with others, we have to define and enforce property rights a contracts. Attribution rights to content distributed over the Internet is going to become more and more important as larger and larger portions of the population move online. If we want to make the highest quality content available at the lowest cost to users, then exclusive rights and privately adaptable rules for allocating them are going to be necessary.

Else the owners of the most valuable content are simply going to keep it offline.

Now Business Leaders Really Have No Excuse

The President himself is taking questions online:

Today, the President invited everyone to use a new feature on WhiteHouse.gov called "Open for Questions" to ask a question about the economy and rate other questions up or down. Then, on Thursday morning, the President will conduct a special online town hall on the economy and answer some of the most popular questions and the event will be streamed on WhiteHouse.gov.

Google's Moderator system is being used.

Bonus Question for Constitutional Lawyers: should the algorithms used here to decide how to moderate be available for public scrutiny?

For earlier posts on closing the feedback loop between top leaders and followers through the Internet see here and here.

This could get interesting

From the President's blog:

I know that some are skeptical about the size and scale of this recovery plan. I understand that skepticism, which is why this recovery plan must and will include unprecedented measures that will allow the American people to hold my Administration accountable for these results. We won’t just throw money at our problems - we’ll invest in what works. Instead of politicians doling out money behind a veil of secrecy, decisions about where we invest will be made public, and informed by independent experts whenever possible. We’ll launch an unprecedented effort to root out waste, inefficiency, and unnecessary spending in our government, and every American will be able to see how and where we spend taxpayer dollars by going to a new website called recovery.gov.

Thank you to Michael A. Gollin

It's been a pleasure to have Michael A. Gollin as a guest here at Broken Symmetry, and I want thank him for sharing his thoughts with readers. Please do make sure to grab a copy of Driving Innovation so that you can benefit more from his insights.

I'm glad that Michael brought up Lincoln again, because I think we agree that his insights on the patent system, although nearly 150 years old to the day, remain a sound basis for our new President to rely upon in considering various patent reforms. In the lecture he gave on Inventions and Discoveries in Jacksonville, Illinois on February 11, 1859, Lincoln declared that:

All nature -- the whole world, material, moral, and intellectual -- is a mine; and, in Adam's day, it was a wholly unexplored mine. Now it is the destined work of Adam's race to develope, by discoveries, inventions, and improvements, the hidden treasures of this mine.

In Driving Innovation, in addition to the metaphor of an engine, you chose to describe the flow of innovation as an "innovation forest." The innovation forest is naturally closer to the ecosystem I had in mind in my reply to your first post. Innovation is not a sterile, individualistic enterprise. Ideas evolve healthiest and fastest in an organic, collaborative and competitive scramble in which new ideas are given resources and protection and older ideas are left alone to be used as a resource directly, or to provide indirectly through their decay into the commons a rich soil for new growth.

In fact, I believe your metaphor is an improvement on Lincoln's mine. From our perspective 150 years on, it is obvious that the natural world (including minerals), are not the bottomless pit that they might have seemed to us in the mid-19th Century. Indeed, there is now a broad consensus that our use of natural resources must be made with an eye toward their soon exhaustion.

How fortunate we are the ideas are more like trees than like minerals! When our minerals and fossil fuels are exhausted, there will be no way for us to replace them in our generation. But with an eye toward the future, and with careful planning, we could someday ensure that we do not use up our living resources before they can be replaced.

Beginning around the time of Lincoln, the incentives of the patent system helped to grow a redwood forest of ideas in the United States. On my view, we have been clearcutting from that forest for the last few years. There is nothing wrong with harvesting part of the bounty of what nature has provided us. But such harvesting should be done with an eye toward leaving enough for the forest to flourish on into the future, so that others may enjoy its benefits after us.

I like the idea that as a patent lawyer, I am a steward of that kind of national and natural treasure. I like your idea of a National Innovation Council, which could assist with determining what parts of the forest are in need of more fertilizer, and which need a brush fire to clear out the dead waste. Perhaps the innovation forest has grown too complex to be regulated under a single set of rules. We would not expect a redwood forest to grow well under the same conditions as a tropical jungle. So why should we expect innovation forests to be any different? I worry, however, that the best forestry experts will not be around to serve on such a council. Only time will tell.

Thanks again for sharing your thoughts.