This is the Second part of the interview with the scientist Professor Kopal about the impact of Big Data on the industry. His lecture about social networks and its impact on business sectors such as telecom and financial industry was brilliant. This is an excellent opportunity to present this scientist and a high-ranked officer of The Ministry of the Interior of the Republic of Croatia. If you missed the First part, you may find it here.
It could be said that crime follows the development of technology. How does Croatian law-enforcement system deal with such challenges? Can you state what kind of knowledge, skills and tools do you use in dealing with security challenges in general?
Development of models based on large data sets for the purpose of forecasting (what might happen?), prediction (what is likely to happen?) and consequently prescription (opportunity cost of decisions) requires advanced analytical skills in terms of technology, but also adjustment of business processes and products to complex nature and new demands imposed by predictive analysis methodology.
In this regard, we might talk about the intention to improve business processes and contribute to the development of predictive analysis by, for example, using software tools as a support to project management (adjustment of analytical processes), visualization tools (product adjustment) and tools which serve for support (automation) to structured and problem-solving analytical techniques (methodology adjustment). Some of (criminal) intelligence challenges are: limited forecasting abilities, inability to predict and prescribe, lack of probability and utility assessments in analysis, lack of weighting for analytical information, inadequate intelligence processes, inadequate process of transformation from data to information, and from information to intelligence, intelligence database as a black hole, multiple data redundancy, inefficient and inadequate exploitation of OSINT, prevailingly relying on descriptive and intuitive analytics, no solution for heuristics, biases, fallacies and the fallibility of mind-sets, no early warning systems management, limited understanding of interconnected risks analysis paralysis, extinct by instinct, expert blindness, analytical paradoxes and failures, "alibi" writing.
The goal that has to be achieved thereby is integration and synergy of separately developed qualitative and quantitative methods from various sciences, control over known and also unknown information, and, apart from 5W-1H model, answering the question So what?, Then what? and Why not?. In this case, apart from qualitative expert judgment and structured analysis, it will be possible to carry out quantitative empirical analysis (data-based computer tools and visualization techniques) and quasi-quantitative analysis (computer-based tools requiring input from experts) by using specific predictive analysis software tools.
In doing so, one must also adequately address the threats and contemporary challenges and risks. There are various risks: economic, ecological, social, geopolitical and technological. However, the true challenge is their interconnectedness. How to realize in time that “…a flap of butterfly’s wings in Brazil can set off a tornado in Texas…”?
Risk can be connected with the likelihood of the occurrence of an adverse event. However, it does not suffice to define the risk solely based on likelihood. The significance of the consequence must also be taken into consideration should such an event occur. On the other hand, uncertainty is a circumstance where there is only awareness of the possibility that an adverse event will occur and no sufficient knowledge of its likelihood.
President Obama said: “…You’re always working with probabilities, and you make a decision, not based on 100% certainty, but with the best information that you’ve got…”
Can you image (hypothetically) the work of an intelligence analyst without any data on the likelihood of the outcome predicted? If there is no likelihood in it, then the likelihood can be defined within the scale ranging from “definitely won’t happen” to “definitely will happen”. These extremes correspond to zero (0%) likelihood and hundred (100%) likelihood.
The issue with calculating likelihood and predicting the consequences is the fact that human brain is not equipped with the “likelihood mode”. We are capable of intuitively and somewhat precisely estimating the height of most objects, their distance or dimension, or make other estimates based on visual perception. However, non-visual and non-tangible estimate of likelihood is often unreliable.
Although it is key to understand the likelihood concept and its use, due to the fact that likelihood intertwines with intelligence analysis, both explicitly and implicitly, the likelihood is a concept that we understand the least and that we find most difficult to use efficiently because likelihood laws are often counterintuitive. We are too often in a phase of uncertainty, whereas we must reach a phase of risk, and after that a phase of risk management.
A phase of uncertainty has to be followed by a phase of risk, and finally a phase of risk management has to be reached, for which human resources are the most important.
The Ministry of the Interior has highly valuable human potential. Those people are our greatest strength. However, in order to tackle contemporary challenges, one has to continuously invest in their education. First and foremost, we would like to enhance our capacities for fighting terrorism, cybercrime and cyber challenges, improve criminal and intelligence analytics and also strengthen the existing capabilities to combat economic and organized crime by digital skills and technologies.
Special attention needs to be paid to the challenges of quantification and interconnectedness of risks and reporting improvement, which includes quantification of descriptive parameters, recommendation development and the opportunity cost assessment in decision-making. It is also of vital importance to improve intelligence analysis with the inclusion of early warning systems, forecasting, prediction and prescription tools and methods, analytical management and the complete mastering of knowledge portfolio. We are also well aware of the impact of digital innovations and disruptive technologies and very keen in our efforts to apply them on a daily basis in the improvement of our security and intelligence system, particularly in areas so important and timely as a systematic address of cyber threats.
Thus, for example, we aim to have in our system those police officers who, other than being familiar with threats arising from today’s digital environment in all areas of life, including crime, also possess certain digital and expert skills which could be efficiently applied in identifying, preventing and solving cybercrime cases. These are persons who can use their high level of technical expertise to efficiently address threats in the digital world. In addition, those are professionals familiar with security practices, politics and rules in general. When it comes to technical skills, apart from possessing digital literacy, they are well familiar with technology and client and server operating system platform, as well as network infrastructure. By nature, they possess a minimum level of knowledge and practical skills related to server system and information and communication networks at system administrator level, and preferably at system engineer level. In addition to such knowledge, they also possess advanced skills in the area of ethical hacking, computer forensics, system security and application penetration testing.
Today, one of the biggest security challenges is terrorism. Several months ago, we heard from Brussels about the upcoming initiatives in improving and enhancing cooperation between the security systems of the EU member-states. How do you view this initiative?
As you have already noted in your previous question, we can see a trend of all sorts of criminal groups adjusting to the latest technical, technological and communication achievements. This is particularly prominent in terrorist organizations, in the area of radicalization as a key element which provides for dissemination and viability of terrorist, radical and extremist ideologies. The evolution of radicalization models is particularly evident when it comes to key radicalization locations. Formerly, radicalization was most common in prisons, religious institutions, citizen associations and similar places. However, in the past few years Internet has taken over as the primary environment and the most common method of radicalization.
In order to prevent online radicalization, in particular radicalization through social networks, as well as the abuse of the cyberspace for the purpose of terrorism, the Ministry of the Interior has implemented numerous platforms and common solutions developed at the EU level.
Following the establishment of the European Internet Referral Unit – EU IRU within Europol, the Republic of Croatia has established the IRU national contact point at the Ministry of the Interior, within the organizational unit of the General Police Directorate responsible for preventing and combating terrorism. Through this national contact point, it informs EU IRU about Internet content that might be related to terrorism.
Likewise, it should be pointed out that the Croatian Ministry of the Interior has set up an electronic form for reporting all Internet content in the Republic of Croatia which may be linked to terrorism. The form can be found at the Ministry’s website. The introduction of this electronic form has allowed all citizens to quickly and anonymously report all Internet content in relation to the Republic of Croatia which might be linked to terrorism. Apart from the above said, it should be pointed out that the Republic of Croatia has been participating in the work of the EU Internet Forum – a European Commission platform gathering the representatives of European Union Member States, bodies and agencies and representatives of key Internet companies, such as Google, Facebook, Twitter, YouTube and others. The Forum examines the possibility of enhancing the existing solutions and implementing new instruments aimed at early detection and automated removal of Internet content related to terrorism, radicalism, violent extremism or radicalization.
Likewise, the Republic of Croatia has joined the new Europol – SIRIUS platform (Shaping Internet Research Investigations Unified System), which will develop new tools for investigating online crimes. However, there are still some challenges which require different approach to possible solutions. I have already publicly spoken and written on this topic on several occasions, so I will briefly repeat what I have said.
In order for Europe to be able to respond to contemporary security challenges, a central European intelligence agency with operational powers has to be set up (EU agencies with analytical capabilities already exist). And this is my expert and not an official opinion. I agree that this will be a difficult task, some might even say impossible. But I think that this is the only right way. Why is it difficult to set up a single security and intelligence agency and/or a single EU information and intelligence system?
On the one hand, due to the Prisoner’s Dilemma, and on the other due to the Stag/Hunt game. But also, because the Tragedy of the Commons. All those terms came from game theory.
In the case of the Prisoner’s Dilemma, individual interest prevails over common interest (to put it simply), and in the Stag Hunt game, the logic "A bird in the hand is worth two in the bush" can be applied (and yet again, to put it very, very simply). One might also say that sometimes national interests prevail over supranational interests. There is also the free rider issue when you do not participate in solving a particular issue (do not bear costs), but you benefit from the resolution of the issue. Lately, you might have come across information in the media in support of a single EU army. International cooperation in security and intelligence structures and information exchange has to be supranational. Furthermore, a single information and intelligence system for raw data/bulk data integration has to be set up (and not just the one for exchanging processed intelligence information).
Europe has taken a step forward in that direction with, for example, the PNR (Passenger Name Record) initiative. Nevertheless, this is not enough. Also public-private partnership in security and intelligence sector has to be further promoted. All this is of critical importance for understanding how risks are interconnected. Microtrends are a special problem. They are based upon the idea that the most powerful forces in our society are the emerging, counterintuitive trends that are shaping tomorrow before us (Penn). However, a microtrend has to be spotted before it reaches 1%, because once it does, it might already become unstoppable. And by that time, it is too late. Unfortunately, I am not convinced that the terrorist threat will diminish. I think that all of us (in Europe and all over the world) will have to adopt new behavioral standards, and our countries new security standards. I am not talking about the dilemma between security or privacy here. Any solution must provide both security and privacy.
Key efforts have to be focused on the 4th quadrant of the matrix of knowledge – “what do we not know that we do not know”, and that might be possible only by applying the data science methods. So, what is data science for me? For me Data Science is like cooking...Data Cooking...Dishes are business problems...Recipes are guidelines...Ingredients are Data Sources...Data Applications... Data Components ...However...
There is a big difference between chefs...and their dishes...You can use same ingredients and make a lousy dish...or an astonishing one...You can add some innovative ingredient...and make your dish even better... It is a challenge for Data Scientists...Data Scientist makes all the difference...Data Scientist is Data Chef... Data Science will enable you to move from deductive reasoning to inductive reasoning, to move toward exploratory data analysis, understanding new relationships, getting insights and analytic paths from the data etc.
Data Science has three tips of expertise: understanding of the reality in which a problem space exists (domain expertise) + the analytical/methodological structure in which problems are examined (analytical/mathematical/statistical expertise) + the environment in which data/intelligence products are created (IT expertise/computer science).
I fully agree with the Booz Allen Hamilton definition of Data Science. Data Science is the art of turning data into actions but also describing Data Science is like trying to describe a sunset – it should be easy, but somehow capturing the words is impossible.
Mastering over hundred analytical techniques and making them work in real-life environment for decades, created your moto: Failure is not an option. If a failure occurs, who is guilty: personnel, equipment or knowledge and skills?
If a failure is not an option, then a failure cannot occur. I am trying to reduce failure to the level of an outlier. On the one hand, I absolutely respect Sun Tzu’s words: “If you know the enemy and know yourself, you need not fear the result of a hundred battles. If you know yourself but not the enemy, for every victory gained you will also suffer a defeat. If you know neither the enemy nor yourself, you will succumb in every battle.”
I aim to apply a multidimensional approach to problem solving by using multiple analytical techniques for the same problem area. This allows me to solve the following intelligence challenges: cognitive limitations and “traps” which make the analysis significantly more difficult; analytic failures that prompted the reviewing of the method in which intelligence analysis is applied and analytic products generated, as well as the wish of decision makers, the end users of analytic products, to ensure more transparent reaching of conclusions and making of predictions.
In doing so, one must also adequately address the threats and contemporary challenges and risks. In this process, human resources are most important. And human resources are people. Therefore, the utmost challenge in mastering security risks is to “defeat oneself”.
Human resources are both the cause and the solution to the problem. Human resources are crucial. I will try to explain this by defining the cause of the problem.
Herbert Simon has introduced the “bounded rationality” concept by which he explained the limitations of cognitive abilities of human mind in interpreting reality that make it unable to deal with the complexity of the real world, and thus prone to creating simplified mental models of reality to be processed at a later stage. People behave rationally within boundaries of their mental models that are not always adjusted to the demands of the real world.
The process of perception links people to their environment and is key to accurate understanding of the world around us. Any precise analysis requires precise perception. Perception is an active and not passive process: it neither documents, nor records reality, it rather creates it. Perception implies understanding of the world around us, and not just awareness of its existence. In other words, perception is a process of deduction by which people create their own versions of reality on the basis of information gathered by their senses.
Therefore, what people in general, especially analysts, perceive, as well as the manner in which they perceive reality is strongly influenced by their previous experiences, education, social norms or society roles, as well as stimuli recorded by sensory organs. Perception is also influenced by its very context, in other words different circumstances give rise to different expectations. All these influences predefine the type of information to which analysts are to pay particular attention and the manner in which obtained information are to be organized and interpreted.
One of the fundamental problems related to perception is that people perceive what they expect to perceive. Patterns of expectation are embedded in such a degree that they influence perception even in situations when people are prepared and aware of the fact that there is information that does not fit their predictions. In other words, the very attempt to be objective does not necessarily ensure accurate perception.
Thus, what an analyst really observes and the manner in which he interprets it partly depends on his patterns of expectation. Analysts have a set of assumptions and expectations about the motivation of people and processes they analyze. Events consistent with those expectations can be easily perceived and processed, while events contradictory to expectations are often ignored or distorted.
The patterns of expectations subconsciously tell the analyst what to look for, what is important and how to interpret what has been seen. Those patterns make up a mental set that predetermines the way the analyst is to deliberate. This mental set is a prism through which we view the world, and objectivity can be accomplished by creating basic assumptions and explicit deliberation which others may try to challenge and analysts question or test their accuracy.
One of the most important characteristics of this mental set is a fact that the mental set can be formed very quickly, yet it is very difficult to change it. This principle is called the perseverance effect. When an analyst creates an image or an expectation about the phenomenon he is studying, this image or expectation will condition the future perception of the studied phenomenon.
An analyst will form his thoughts and develop his hypotheses on the basis of his own observation. The more certain an analyst is of his initial perception, the greater its influence on all later perception. New evidence will be assimilated into this initial image as long as it is not significantly contradictory to his initial perception and this contradiction becomes so evident that it imposes itself onto the analyst’s consciousness. The initial, although inaccurate, perception will resist change as the amount of information necessary to reject the hypothesis is significantly larger than the amount necessary to make the initial interpretation. Another cause of many analytical problems is limitation of the so-called “working memory” of human mind. George Miller came to the conclusion that people can hold 7 items in their mind at a time (plus or minus two). The fact that it is difficult for human mind to fully grasp complex problems makes the decision-making more difficult. For example, first we will come up with arguments in favor, then those against, but we will not be able to simultaneously hold all of them in our mind in order to compare and analyze them. The methodology recommended for surpassing these obstacles is problem externalization and decomposition. It presupposes “shifting” the problem from the mind by writing it down on a piece of paper in a simplified broken down form into main elements and their correlations.
This is the very principle on which structured analytical techniques are based. What all these techniques have in common is that they include breaking down the problem into its integral parts and development of simple models which demonstrate the relationship of these parts and the whole. When we analyze a small part of the problem, the model helps us not to lose sight of the problem as a whole. A simple model of an analytical problem enables assimilation of new information into a long-term memory and it offers structure with which we can connect parts of information. The model defines categories for storing information into memory and their retrieval when necessary.
People tend to avoid analytical structures because structuring is contrary to the way human mind functions. We are susceptible to prejudice and assumptions and we instinctively rely on them. Prejudice is a subconscious belief which conditions, governs and motivates our behavior. Generally speaking, prejudice can be valuable as it provides mental shortcuts for extremely fast processing of new information, and our everyday functioning would not be possible without it. However, the speed of this process and the fact that it is subconscious, and beyond our control, have an adverse effect as they increase and confirm our prejudice at the expense of the truth. The reason we might be confused by prejudice lies in the fact that our mind does not thoroughly evaluate the logic of each new received piece of information. Instead, our mind uses a shortcut based on a pattern; it functions analogically, and not logically.
True objectivity is also very rare, first and foremost due to the instinctively subjective manner in which human brain is programmed to function. Analyses often focus on solutions which we intuitively prefer, whereby not enough attention is paid to alternative solutions. Inability to fully take into consideration alternative solutions is the most common cause of poor and incomplete analysis, which often results in wrong decisions and ultimately business failure.
The solution which we intuitively prefer is the first one which we find satisfying. This phenomenon is called “satisficing” (a combination of words satisfy and suffice). This coined word refers to the realization that decision makers most often prefer to accept the solutions that are temporarily satisfying, rather than looking for a better solution which would be possible if “rational model” were used.
In other words, we must learn to become receptive towards new realizations. Structured analytical techniques help identify and break down restrictive mental sets which prevent us from becoming more receptive to new realizations and accepting alternative solutions. These techniques provide a substitute for the limitations of human mind in analyzing complex problems that often include ambiguous and obscure information, a large number of players and changing circumstances.
It should be noted that structuring is not a substitute for deliberation but a means which makes deliberation easier and more effective. We need an evolution of decision making toward Data-Driven Decision-making. In the research "Exploring the agenda for big decisions in 2014-15 (PwC)" one question was: Which of the following inputs did you place the most reliance on for your last big decision? 30% said that they put most reliance on their own intuition an experience and almost 30% more on the advice or experience of other internally. See the problem? Where is Data-Driven Decision-making in this? And what about cognitive and other implications (biases, heuristics etc.) if we are relying on everything but the data? This concise form of deliberation is only one of the reasons why structured approach has to be used in intelligence analysis. This is just one of the dimensions that I use in my work on a daily basis.
What skills are vital for dealing with security challenges of the 21st century?
The hardest skills to find are those that can't be performed by machines: problem solving skills, creativity and innovation, adaptability, leadership, emotional intelligence (20th CEO Survey (2016. - PwC)). Also, according to the World Economic Forum (Future of Jobs Report) the top 3 skills in 2020 will be complex problem solving, critical thinking and creativity. Because of that, we (Ms Korkut and myself) have designed 3 courses for the university college Algebra where we are also teaching: Design Thinking 1 - creativity and critical thinking, Design Thinking 2 - problem solving skills and Design Thinking 3 - business and competitive analysis.
We have also designed game theory course (also for University College Effectus) and social network analysis course. Furthermore, we have designed the data science and digital marketing master studies. And the first books about those subjects we have published in 2011., before it become "mainstream". But we are also trying to make a difference in security area. We are teaching 4 levels of intelligence analysis: basic, advanced, expert and strategic level. Some of techniques in the Basic Intelligence Analysis course are: Cognitive Limitations, Link/Network Analysis, Inference Development, Commodity Flow Charting, Event Charting, Activity Charting, Analysis of Competing Hypotheses, Collection Plan etc.
Some of techniques in the Advanced Intelligence Analysis course are: Causal Flow Diagramming, Weighted Ranking, Probability Tree, Utility Tree, Advanced Utility Analysis, Assessment of Inferences, Concealed Income Analysis, Social Network Analysis, Social Physics, Briefings and Presentation Skills etc. Some of techniques in the Expert Intelligence Analysis course are: Convergent & Divergent Thinking, Structured Brainstorming, Morphological Analysis, Quadrant Crunching, De Bono 6 Hats, Issue Redefinition – Problem Restatement, Mind Mapping, Scenarios Analysis, Indicators & Indicators Validator, Hypothesis Generation Techniques, Alternative Futures etc.
Some of techniques in the Strategic Intelligence Analysis course are: Strategic Intelligence, Reasoning and Logic, Biases, Fallacies, Critical Thinking, Terms of Reference - ToR, Data Collection Plans for Strategic Projects, PESTEL, Extended and Advanced TOWS analysis, Threat Measurement Techniques for Strategic Intelligence Analysis etc. Last month I have conducted advanced integrated criminal intelligence course for 80 police and intelligence officers.
Can we say that Big Data is a part of digital forensics? How did you come to the idea of becoming a scientist and exploring this kind of science?
A dilemma that I had and resolved myself many years ago was: professional researcher vs. researching professional. Which of the two do I want to be? Which one makes me more fulfilled and a better person? The answer was: a researching professional. Thus, my personal choice was to be an applied data scientist. I am a lecturer at university colleges and faculties. I research and learn, and I apply all that I have researched and learned. Then I approach all the challenges I have come across in practice as a scientist and try to solve and reapply them accordingly. Hence, I led two lives for years, from 8 a.m. to 4 p.m. within the security system, and from 5 p.m. onwards as a member of academic and scientific community. This allowed me to implement the latest achievements, the latest technologies and the latest scientific findings into the security system. Currently I am State Secretary at the Ministry of the Interior. Although this is politician position I am nevertheless a data scientist. In one paper about data science in politics I wrote "...The first part of the formula for success in politics is to make sure data scientists are not politicians. The other part is on the side of politicians..."
So, I am data scientist in the politics but I am not a politician. Lately I have publicly spoken on the topic of security and security challenges on many occasions with the aim of defining the direction that Croatia and the EU have to follow to overcome security challenges of the future. Experts of various profiles have also provided their expert and professional opinion and comments on numerous, unfortunately sad and painful events – terrorist acts. Thus, "phenomenological" experts focused on the phenomenology of terrorism (causes and consequences, history and evolution of terrorism, etc.), whereas “methodological” experts focused on the methodology of contemporary terrorism (methods and means, tactics and strategies, etc.). Although they truly possess top expertise which is at the same time an essential component for the efficient fight against terrorism, the characteristic they have in common is that they mostly analyze the past or maybe the present. They do, of course, make predictions about the future. However, these predictions are almost exclusively based on description, intuition and expertise.
But how can one conceive the inconceivable? How “not to be surprised” in the upcoming future? Well, for example, by thinking counterintuitively. And thinking counterintuitively is extremely difficult. Excellent examples of the counterintuitive were provided by Bruce Bueno de Mesquita in his outstanding book “The Predictioneers Game”. By the way, this is the same de Mesquita that History Channel made a 90-minute show about, called “The Next Nostradamus”. This is also the same de Mesquita who has made predictions with over 90% precision in a five-year period, and who challenged by CIA analysts, proved to be better in predictions than the best among them. De Mesquita is using game theory and mathematics in order to be able to solve such challenges. De Mesquita has also developed mathematical models for the North Korean disarmament negotiations and the Middle East negotiation process. And making prediction is not an easy task. The research conducted by Philip Tetlock is the best indicator of how difficult it is to make predictions. In his book “Expert Political Judgement”, he presented the results of a long-term study conducted in the period from 1985 to 2003 on the accuracy of prediction, in which he examined 82,361 forecasts made by 284 experts. These are some of the results of the study: 1. Experts are not significantly more efficient than non-experts; 2. Statistical regression models have achieved better results than human experts; 3. Experts are less efficient than dart-throwing monkeys, and 4. Too much knowledge can have an adverse effect on the accuracy of prediction.
What I am interested in and what I do is quantitative and structural analysis of the future, which leads us to the third type of experts – problem solvers (these days known as data scientists). Some of the methods used thereby are alternative future analysis, multiple scenarios generation, multiple hypotheses generator, analysis of competing hypotheses, indicator generator and validator, quadrant analysis, advanced utility analysis, game theory and many others.
The art of analyzing is the art of asking questions, such as the following: What must be done at national and international level for all of us to be more successful in combating terrorism?; How to recognize a potential crisis?; How to recognize a latent crisis?; How to build an efficient early warning system?; How to shift from a reactive approach to a proactive approach?; What is analytic management and how to implement it?; How to achieve integration and synergy of separately developed qualitative and quantitative methods from various science fields in a security and intelligence system?; How to implement data science approach into a security and intelligence system?; How to find answers to key intelligence questions: what?, so what?, then what? and why not?; How to shift from reporting, analysis and monitoring to forecasting, prediction and prescription?; How to fully master the matrix of knowledge: what do we know that we know; what do we know that we do not know; what do we not know that we know; what do we not know that we do not know?, etc. For example, the focus of my scientific work has lately been social physics.
New, all-present digital data on all aspects of human life have become readily available. By using these data to build a predictive computational theory of human behavior, we can hope to engineer better and more credible communities and societies. MIT Professor Alex “Sandy” Pentland is the author of a popular book entitled Social Physics: How Good Ideas Spread – The Lessons from a New Science, which won the 2014 Best Business Book Award. In his book, he lists a number of specific ways that methods of social physics can be applied to the way that organizations conduct their business, but also to everyday, social relationships. Pentland asks the question: How can we create organizations and governments that are cooperative, productive and creative? He claims that these are the questions of social physics that are especially important right now because of increased global competition, global challenges and frequent government failures.
Pentland’s book ignited a spark in me and I completed the MIT Big Data and Social Analytics course. What is social physics actually? Social physics is a quantitative social science that describes reliable, mathematical connections between information and flow of ideas on the one hand, and human behavior on the other. It helps us understand how the ideas flow from one person to another through the mechanism of social learning and how this flow of ideas ends up shaping the norms, productivity and creative output of companies, cities and societies. The application of methods of social physics, for example by tuning communication networks, has proven to enhance productivity and reliability in decision-making. And decision-making is precisely the field of social physics that I am currently working on.