Search This Blog

Tuesday, November 13, 2018

Neuroeconomics

From Wikipedia, the free encyclopedia

Neuroeconomics is an interdisciplinary field that seeks to explain human decision making, the ability to process multiple alternatives and to follow a course of action. It studies how economic behavior can shape our understanding of the brain, and how neuroscientific discoveries can constrain and guide models of economics.

It combines research from neuroscience, experimental and behavioral economics, and cognitive and social psychology. As research into decision-making behavior becomes increasingly computational, it has also incorporated new approaches from theoretical biology, computer science, and mathematics. Neuroeconomics studies decision making by using a combination of tools from these fields so as to avoid the shortcomings that arise from a single-perspective approach. In mainstream economics, expected utility (EU) and the concept of rational agents are still being used. Many economic behaviors are not fully explained by these models, such as heuristics and framing.

Behavioral economics emerged to account for these anomalies by integrating social, cognitive, and emotional factors in understanding economic decisions. Neuroeconomics adds another layer by using neuroscientific methods in understanding the interplay between economic behavior and neural mechanisms. By using tools from various fields, some scholars claim that neuroeconomics offers a more integrative way of understanding decision making.

Introduction

The field of decision making is largely concerned with the processes by which individuals make a single choice from among many options. These processes are generally assumed to proceed in a logical manner such that the decision itself is largely independent of context. Different options are first translated into a common currency, such as monetary value, and are then compared to one another and the option with the largest overall utility value is the one that should be chosen. While there has been support for this economic view of decision making, there are also situations where the assumptions of optimal decision making seem to be violated.

The field of neuroeconomics arose out of this controversy. By determining which brain areas are active in which types of decision processes, neuroeconomists hope to better understand the nature of what seem to be suboptimal and illogical decisions. While most of these scientists are using human subjects in this research, others are using animal models where studies can be more tightly controlled and the assumptions of the economic model can be tested directly.

For example, Padoa-Schioppa & Assad tracked the firing rates of individual neurons in the monkey orbitofrontal cortex while the animals chose between two kinds of juice. The firing rate of the neurons was directly correlated with the utility of the food items and did not differ when other types of food were offered. This suggests that, in accordance with the economic theory of decision making, neurons are directly comparing some form of utility across different options and choosing the one with the higher value. Similarly, a common measure of prefrontal cortex dysfunction, the FrSBe, is correlated with multiple different measures of economic attitudes and behavior, supporting the idea that brain activation can display important aspects of the decision process.

Neuroeconomics studies the neurobiological along with the computational bases of decision-making. A framework of basic computations which may be applied to Neuroeconomics studies is proposed by A. Rangel, C. Camerer, and P. R. Montague. It divides the process of decision making into five stages implemented by a subject. First, the representation of the problem is computed. This includes analysis of internal states, external states and potential course of action. Second, values are assigned to potential actions. Third, based on the valuations, one of the actions is selected. Fourth, the subject evaluates how desirable the outcome is. Final stage, learning, includes updating all of the above processes in order to improve future decisions.

Major research areas

Decision making under risk and uncertainty

Most of our decisions are made under some conditions of risk. Decision sciences such as psychology and economics usually define risk as the uncertainty about several possible outcomes when the probability of each is known. Utility maximization, first proposed by Daniel Bernoulli in 1738, is used to explain decision making under risk. The theory assumes that humans are rational and will assess options based on the expected utility they will gain from each.

Research and experience uncovered a wide range of expected utility anomalies and common patterns of behavior that are inconsistent with the principle of utility maximization – for example, the tendency to overweight small probabilities and underweight large ones. Daniel Kahneman and Amos Tversky proposed prospect theory to encompass these observations and offers an alternative model.

There seem to be multiple brain areas involved in dealing with situations of uncertainty. In tasks requiring individuals to make predictions when there is some degree of uncertainty about the outcome, there is an increase in activity in area BA8 of the frontomedian cortex as well as a more generalized increase in activity of the mesial prefrontal cortex and the frontoparietal cortex. The prefrontal cortex is generally involved in all reasoning and understanding, so these particular areas may be specifically involved in determining the best course of action when not all relevant information is available.

In situations that involve known risk rather than uncertainty, the insular cortex seems to be highly active. For example, when subjects played a 'double or nothing' game in which they could either stop the game and keep accumulated winnings or take a risky option resulting in either a complete loss or doubling of winnings, activation of the right insula increased when individuals took the gamble. It is hypothesized that the main role of the insular cortex in risky decision making is to simulate potential negative consequences of taking a gamble.

In addition to the importance of specific brain areas to the decision process, there is also evidence that the neurotransmitter dopamine may transmit information about uncertainty throughout the cortex. Dopaminergic neurons are strongly involved in the reward process and become highly active after an unexpected reward occurs. In monkeys, the level of dopaminergic activity is highly correlated with the level of uncertainty such that the activity increases with uncertainty. Furthermore, rats with lesions to the nucleus accumbens, which is an important part of the dopamine reward pathway through the brain, are far more risk averse than normal rats. This suggests that dopamine may be an important mediator of risky behavior.

Individual level of risk aversion among humans is influenced by testosterone concentration. There are studies exhibiting correlation between the choice of a risky career (financial trading, business) and testosterone exposure. As markers for prenatal ("organizational") testosterone transfer, 2D:4D ratio and Baron-Cohen test can be used, whereas circulating ("activational") testosterone can be measured directly via its concentration in saliva. It appears that lower digit ratio preconditions a longer and more successful career in financial trading, especially for markets of high volatility. In addition, daily achievements of traders with lower digit ratio are more sensitive to circulating testosterone. A long-term study of risk aversion and risky career choice was conducted for a representative group of MBA students. It revealed that females are in average more risk averse, but the difference between genders vanishes for low organizational and activational testosterone exposure leading to risk-averse behaviour. Students with high salivary testosterone concentration and low digit ratio, disregarding the gender, tend to choose risky career in finance (e.g. trading or investment banking).

Serial and functionally localized model vs distributed, hierarchical model

In 2017 March, Laurence T. Hunt and Benjamin Y. Hayden argued an alternative viewpoint of the mechanistic model to explain how we evaluate options and choose the best course of action. Many accounts of reward-based choice argue for distinct component processes that are serial and functionally localized. The component processes typically include the evaluation of options, the comparison of option values in the absence of any other factors, the selection of an appropriate action plan and the monitoring of the outcome of the choice. They emphasized how several features of neuroanatomy may support the implementation of choice, including mutual inhibition in recurrent neural networks and the hierarchical organization of timescales for information processing across the cortex.

Loss aversion

One aspect of human decision making is a strong aversion to potential loss. For example, the cost of losing a specific amount of money is higher than the value of gaining the same amount of money. One of the main controversies in understanding loss aversion is whether the process is driven by a single neural system that directly compares options and decides between them or whether there are competing systems, one responsible for a reasoned comparison among options and another more impulsive and emotional system driven by an aversion to potentially negative outcomes.

While one study found no evidence for an increase in activation in areas related to negative emotional reactions in response to loss aversion another found that individuals with damaged amygdalas had a lack of loss aversion even though they had normal levels of general risk aversion, suggesting that the behavior was specific to potential losses. These conflicting studies suggest that more research needs to be done to determine whether there are areas in the brain that respond specifically to potential loss or whether loss aversion is the byproduct of more general reasoning processes.

Another controversy in loss aversion research is whether losses are actually experienced more negatively than equivalent gains or merely predicted to be more painful but actually experienced equivalently. Neuroeconomic research has attempted to distinguish between these hypotheses by measuring different physiological changes in response to both loss and gain. Studies have found that skin conductance, pupil dilation and heart rate are all higher in response to monetary loss than to equivalent gain. All three measures are involved in stress responses, so it seems that losing a particular amount of money is experienced more strongly than gaining the same amount.

Intertemporal choice

In addition to risk preference, another central concept in economics is intertemporal choices which are decisions that involve costs and benefits that are distributed over time. Intertemporal choice research studies the expected utility that humans assign to events occurring at different times. The dominant model in economics which explains it is discounted utility (DU). DU assumes that humans have consistent time preference and will assign value to events regardless of when they occur. Similar to EU in explaining risky decision making, DU is inadequate in explaining intertemporal choice.

For example, DU assumes that people who value a bar of candy today more than 2 bars tomorrow, will also value 1 bar received 100 days from now more than 2 bars received after 101 days. There is strong evidence against this last part in both humans and animals, and hyperbolic discounting has been proposed as an alternative model. Under this model, valuations fall very rapidly for small delay periods, but then fall slowly for longer delay periods. This better explains why most people who would choose 1 candy bar now over 2 candy bars tomorrow, would, in fact, choose 2 candy bars received after 101 days rather than the 1 candy bar received after 100 days which EU assumes.

Neuroeconomic research in intertemporal choice is largely aimed at understanding what mediates observed behaviors such as future discounting and impulsively choosing smaller sooner rather than larger later rewards. The process of choosing between immediate and delayed rewards seems to be mediated by an interaction between two brain areas. In choices involving both primary (fruit juice) and secondary rewards (money), the limbic system is highly active when choosing the immediate reward while the lateral prefrontal cortex was equally active when making either choice. Furthermore, the ratio of limbic to cortex activity decreased as a function of the amount of time until reward. This suggests that the limbic system, which forms part of the dopamine reward pathway, is most involved in making impulsive decisions while the cortex is responsible for the more general aspects of the intertemporal decision process.

The neurotransmitter serotonin seems to play an important role in modulating future discounting. In rats, reducing serotonin levels increases future discounting while not affecting decision making under uncertainty. It seems, then, that while the dopamine system is involved in probabilistic uncertainty, serotonin may be responsible for temporal uncertainty since delayed reward involves a potentially uncertain future. In addition to neurotransmitters, intertemporal choice is also modulated by hormones in the brain. In humans, a reduction in cortisol, released by the hypothalamus in response to stress, is correlated with a higher degree of impulsivity in intertemporal choice tasks. Drug addicts tend to have lower levels of cortisol than the general population, which may explain why they seem to discount the future negative effects of taking drugs and opt for the immediate positive reward.

Social decision making

While most research on decision making tends to focus on individuals making choices outside of a social context, it is also important to consider decisions that involve social interactions. The types of situations that decision theorists study are as diverse as altruism, cooperation, punishment, and retribution. One of the most frequently utilized tasks in social decision making is the prisoner's dilemma.

In this situation, the payoff for a particular choice is dependent not only on the decision of the individual but also on that of another individual playing the game. An individual can choose to either cooperate with his partner or defect against the partner. Over the course of a typical game, individuals tend to prefer mutual cooperation even though defection would lead to a higher overall payout. This suggests that individuals are motivated not only by monetary gains but also by some reward derived from cooperating in social situations.

This idea is supported by neural imaging studies demonstrating a high degree of activation in the ventral striatum when individuals cooperate with another person but that this is not the case when people play the same prisoner's dilemma against a computer. The ventral striatum is part of the reward pathway, so this research suggests that there may be areas of the reward system that are activated specifically when cooperating in social situations. Further support for this idea comes from research demonstrating that activation in the striatum and the ventral tegmental area show similar patterns of activation when receiving money and when donating money to charity. In both cases, the level of activation increases as the amount of money increases, suggesting that both giving and receiving money results in neural reward.

An important aspect of social interactions such as the prisoner's dilemma is trust. The likelihood of one individual cooperating with another is directly related to how much the first individual trusts the second to cooperate; if the other individual is expected to defect, there is no reason to cooperate with them. Trust behavior seems to be related to the presence of oxytocin, a hormone involved in maternal behavior and pair bonding in many species. When oxytocin levels were increased in humans, they were more trusting of other individuals than a control group even though their overall levels of risk-taking were unaffected suggesting that oxytocin is specifically implicated in the social aspects of risk taking.

One more important paradigm for neuroeconomic studies is ultimatum game. In this game Player 1 gets a sum of money and makes decision how much he wants to split with Player 2. Player 2 either accepts or rejects the offer. If he accepts both players get the amount as proposed by Player 1, if he rejects nobody gets anything. Rational strategy for Player 2 would be to accept any offer because it has more value than zero. However, it has been shown that people often disregard offers that they consider as unfair. Neuroimaging studies indicated several brain regions that are activated in response to unfairness in ultimatum game. They include bilateral mid-anterior insula, anterior cingulate cortex (ACC), medial supplementary motor area (SMA), cerebellum and right dorsolateral prefrontal cortex (DLPFC). It has been shown that low-frequency repetitive transcranial magnetic stimulation of DLPFC increases the likelihood of accepting unfair offers in the ultimatum game.

Another issue in the field of neuroeconomics is represented by role of reputation acquisition in social decision making. Social exchange theory claims that prosocial behavior originates from the intention to maximize social rewards and minimize social costs. In this case approval from others may be viewed as a significant positive reinforcer - i.e., a reward. Neuroimaging studies have provided evidence supporting this idea – it was shown that processing of social rewards activates striatum, especially left putamen and left caudate nucleus, in the same fashion these areas are activated during the processing of monetary rewards. These findings also support so-called “common neural currency” idea, which assumes existence of shared neural basis for processing of different types of reward.

Methodology

Behavioral economics experiments record the subject's decision over various design parameters and use the data to generate formal models that predict performance. Neuroeconomics extends this approach by adding observation of the nervous system to the set of explanatory variables. The goal of neuroeconomics is to inform the creation and contribute another layer of data to the testable hypotheses of these models.

Furthermore, neuroeconomic research is being used to understand and explain aspects of human behavior that do not conform to traditional economic models. While these behavior patterns are generally dismissed as 'fallacious' or 'illogical' by economists, neuroeconomic researchers are trying to determine the biological reasons for these behaviors. By using this approach, we may be able to find valid reasons for the presence of these seemingly sub-optimal behaviors.

Neurobiological research techniques

There are several different techniques that can be utilized to understand the biological basis of economic behavior. Neural imaging is used in human subjects to determine which areas of the brain are most active during particular tasks. Some of these techniques, such as fMRI or PET are best suited to giving detailed pictures of the brain which can give information about specific structures involved in a task. Other techniques, such as ERP (event-related potentials) and oscillatory brain activity are used to gain detailed knowledge of the time course of events within a more general area of the brain.

In addition to studying areas of the brain, some studies are aimed at understanding the functions of different brain chemicals in relation to behavior. This can be done by either correlating existing chemical levels with different behavior patterns or by changing the amount of the chemical in the brain and noting any resulting behavioral changes. For example, the neurotransmitter serotonin seems to be involved in making decisions involving intertemporal choice while dopamine is utilized when individuals make judgments involving uncertainty. Furthermore, artificially increasing oxytocin levels increases trust behavior in humans while individuals with higher cortisol levels tend to be more impulsive and exhibit more future discounting.

In addition to studying the behavior of normal individuals in decision making tasks, some research involves comparing the behavior of normal individuals to that of others with damage to areas of the brain expected to be involved in certain behaviors. In humans, this means finding individuals with specific types of neural impairment. For example, people with amygdala damage seem to exhibit less loss aversion than normal controls. Also, scores from a survey measuring correlates of prefrontal cortex dysfunction are correlated with general economic attitudes.

Previous studies investigated the behavioral patterns of patients with psychiatric disorders, such as schizophrenia, autism, depression, or addiction, to get the insights of their pathophysiology. In animal studies, highly controlled experiments can get more specific information about the importance of brain areas to economic behavior. This can involve either lesioning entire brain areas and measuring resulting behavior changes or using electrodes to measure the firing of individual neurons in response to particular stimuli.

Notable theorists

Experiments

In a typical behavioral economics experiment, a subject is asked to make a series of economic decisions. For example, a subject may be asked whether they prefer to have 45 cents or a gamble with a 50% chance to win one dollar. The experimenter will then measure different variables in order to determine what is going on in the subject's brain as they make the decision. Some authors have demonstrated that neuroeconomics may be useful not only to describe experiments involving rewarding but may also be applied in order to describe the psychological behavior of common psychiatric syndromes involving addiction as well as delusion.

Criticism

Glenn W. Harris and Emanuel Donchin have criticized the emerging field. Example of critics have been that it is "a field that oversells itself"; or that neuroeconomic studies "misunderstand and underestimate traditional economic models". A critical argument of traditional economists against the neuroeconomic approach, is that the use of non-choice data, such as response times, eye-tracking and neural signals that people generate during decision making, should be excluded from any economic analysis.

Neuromarketing

Neuromarketing is a distinct discipline closely related to neuroeconomics. While neuroeconomics has more academic aims, since it studies the basic mechanisms of decision-making, neuromarketing is an applied field which uses neuroimaging tools for market investigations.

Collective narcissism

From Wikipedia, the free encyclopedia

Collective narcissism (or group narcissism) extends the concept of individual narcissism onto the social level of self. It is a tendency to exaggerate the positive image and importance of a group the individual belongs to – i.e. the ingroup. While the classic definition of narcissism focuses on the individual, collective narcissism asserts that one can have a similar excessively high opinion of a group, and that a group can function as a narcissistic entity. Collective narcissism is related to ethnocentrism. However, ethnocentrism primarily focuses on self-centeredness at an ethnic or cultural level, while collective narcissism is extended to any type of ingroup, beyond just cultures and ethnicities. While ethnocentrism is an assertion of the ingroup's supremacy, collective narcissism is a self-defensive tendency to invest unfulfilled self-entitlement into a belief about ingroup's uniqueness and greatness. Thus, the ingroup is expected to become a vehicle of actualisation of frustrated self-entitlement. When applied to a national group, collective narcissism is similar to nationalism: a desire for national supremacy. Positive overlap between ingroup satisfaction and collective narcissism suppresses collective narcissistic intergroup hostility.

Development of the concept

In Sigmund Freud's 1922 study Group Psychology and the Analysis of the Ego, he noted how every little canton looks down upon the others with contempt, as an instance of what would later to be termed Freud's theory of collective narcissism. Wilhelm Reich and Isaiah Berlin explored what the latter called the rise of modern national narcissism: the self-adoration of peoples. "Group narcissism" is described in a 1973 book entitled The Anatomy of Human Destructiveness by psychologist Erich Fromm. In the 1990s, Pierre Bourdieu wrote of a sort of collective narcissism affecting intellectual groups, inclining them to turn a complacent gaze on themselves. Noting how people's desire to see their own groups as better than other groups can lead to intergroup bias, Henri Tajfel approached the same phenomena in the seventies and eighties, so as to create social identity theory, which argues that people's motivation to obtain positive self-esteem from their group memberships is one driving-force behind in-group bias. The term "collective narcissism" was highlighted anew by researcher Agnieszka Golec de Zavala who created the Collective Narcissism Scale and developed research on intergroup and political consequences of collective narcissism. People who score high on the Collective Narcissists Scale agree that their group's importance and worth are not sufficiently recognised by others and that their group deserves special treatment. They insist that their group must obtain special recognition and respect.

The Scale was modelled on the Narcissistic Personality Inventory. However, collective and individual narcissism are modestly correlated. Only collective narcissism predicts intergroup behaviours and attitudes. Collective narcissism is related to vulnerable narcissism (individual narcissism manifesting as distrustful and neurotic interpersonal style), rather than grandiose narcissism (individual narcissism manifesting as exceedingly self-aggrandising interpersonal style) and to low self-esteem. This is in line with theorising of Theodore Adorno who proposed that collective narcissism motivated support for the Nazi politics in Germany and was a response to undermined sense of self-worth.

Characteristics and consequences

Collective narcissism is characterized by the members of a group holding an inflated view of their ingroup which requires external validation. Collective narcissism can be exhibited by an individual on behalf of any social group or by a group as a whole. Research participants found that they could apply statements of the Collective Narcissism Scale to various groups: national, ethnic, religious, ideological, political, students of the same university, fans of the same football team, professional groups and organizations Collectively narcissistic groups require external validation, just as individual narcissists do. Organizations and groups who exhibit this behavior typically try to protect their identities through rewarding group-building behavior (this is positive reinforcement).

Collective narcissism predicts retaliatory hostility to past, present, actual and imagined offences to the ingroup and negative attitudes towards groups perceived as threatening. It predicts constant feeling threatened in intergroup situations that require a stretch of imagination to be perceived as insulting or threatening. For example, in Turkey, collective narcissists felt humiliated by the Turkish wait to be admitted to the European Union. After a transgression as petty as a joke made by a Polish celebrity about the country's government, Polish collective narcissists threatened physical punishment and openly rejoiced in the misfortunes of their "offender". Collective narcissism predicts conspiracy thinking about secretive malevolent actions of outgroups.

Golec de Zavala et al. state some parallels between individual and collective narcissism:

Individual/Collective Narcissism Equivalencies
Individual Collective
I wish people would recognize my authority I wish other people would recognize the authority of my group
I have natural talent for influencing people My group has all predispositions to influence others
If I ruled the world it would be a much better place If my group ruled the world it would be a much better place
I am an extraordinary person My group is extraordinary
I like to be the center of attention I like when my group is the center of attention
I will never be satisfied until I get what I deserve I will never be satisfied until my group gets all that it deserves
I insist upon getting the respect that is due to me I insist upon my group getting the respect that is due to it
I want to amount to something in the eyes of the world I want my group to amount to something in the eyes of the world
People never give me enough recognition for the things I've done Not many people seem to understand the full importance of my group

Collective vs. individual

There are several connections, and intricate relationships between collective and individual narcissism, or between individual narcissism stemming from group identities or activities. No single relationship between groups and individuals, however, is conclusive or universally applicable. In some cases, collective narcissism is an individual's idealization of the ingroup to which it belongs, while in another the idealization of the group takes place at a more group-level, rather than an instillation within each individual member of the group. In some cases, one might project the idealization of himself onto his group, while in another case, the development of individual-narcissism might stem from being associated with a prestigious, accomplished, or extraordinary group.

An example of the first case listed above is that of national identity. One might feel a great sense of love and respect for one's nation, flag, people, city, or governmental systems as a result of a collectively narcissistic perspective. It must be remembered that these feelings are not explicitly the result of collective narcissism, and that collective narcissism is not explicitly the cause of patriotism, or any other group-identifying expression. However, glorification of one's group (such as a nation) can be seen in some cases as a manifestation of collective narcissism.

In the case where the idealization of self is projected onto ones group, group-level narcissism tends to be less binding than in other cases. Typically in this situation the individual—already individually narcissistic—uses a group to enhance his own self-perceived quality, and by identifying positively with the group and actively building it up, the narcissist is enhancing simultaneously both his own self-worth, and his group's worth. However, because the link tends to be weaker, individual narcissists seeking to raise themselves up through a group will typically dissociate themselves from a group they feel is damaging to their image, or that is not improving proportionally to the amount of support they are investing in the group.

Involvement in one's group has also been shown to be a factor in the level of collective narcissism exhibited by members of a group. Typically a more involved member of a group is more likely to exhibit a higher opinion of the group. This results from an increased affinity for the group as one becomes more involved, as well as a sense of investment or contribution to the success of the group. Also, another perspective asserts that individual narcissism is related to collective narcissism exhibited by individual group members. Personal narcissists, seeing their group as a defining extension of themselves, will defend their group (collective narcissism) more avidly than a non-narcissist, to preserve their own perceived social standing along with their group's. In this vein, a problem is presented; for while an individual narcissist will be heroic in defending his or her ingroup during intergroup conflicts, he or she may be a larger burden on the ingroup in intragroup situations by demanding admiration, and exhibiting more selfish behavior on the intragroup level—individual narcissism.

Conversely, another relationship between collective narcissism and the individual can be established with individuals who have a low or damaged ego investing their image in the well-being of their group, which bears strong resemblance to the "ideal-hungry" followers in the charismatic leader-follower relationship. As discussed, these ego-damaged group-investors seek solace in belonging to a group; however, a charismatic, strong leader is not always requisite for someone weak to feel strength by building up a narcissistic opinion of their own group.

The charismatic leader-follower relationship

Another sub-concept encompassed by collective narcissism is that of the "Charismatic Leader-Follower Relationship" theorized by political psychologist Jerrold Post. Post takes the view that collective narcissism is exhibited as a collection of individual narcissists, and discusses how this type of relationship emerges when a narcissistic charismatic leader, appeals to narcissistic "ideal-hungry" followers.

An important characteristic of the leader follower-relationship are the manifestations of narcissism by both the leader and follower of a group. Within this relationship there are two categories of narcissists: the mirror-hungry narcissist, and the ideal-hungry narcissist—the leader and the followers respectively. The mirror-hungry personality typically seeks a continuous flow of admiration and respect from his followers. Conversely, the ideal-hungry narcissist takes comfort in the charisma and confidence of his mirror-hungry leader. The relationship is somewhat symbiotic; for while the followers provide the continuous admiration needed by the mirror-hungry leader, the leader's charisma provides the followers with the sense of security and purpose that their ideal-hungry narcissism seeks. Fundamentally both the leader and the followers exhibit strong collectively narcissistic sentiments—both parties are seeking greater justification and reason to love their group as much as possible.

Perhaps the most significant example of this phenomenon would be that of Nazi Germany. Adolf Hitler's charisma and polarizing speeches satisfied the German people's hunger for a strong leader. Hitler's speeches were characterized by their emphasis on "strength"—referring to Germany—and "weakness"—referring to the Jewish people. Some have even described Hitler's speeches as "hypnotic"—even to non-German speakers—and his rallies as "watching hypnosis on large scale". Hitler's charisma convinced the German people to believe that they were not weak, and that by destroying the perceived weakness from among them (the Jews), they would be enhancing their own strength—satisfying their ideal-hungry desire for strength, and pleasing their mirror-hungry charismatic leader.

Intergroup aggression

Collective narcissism has been shown to be a factor in intergroup aggression and bias. Primary components of collectively narcissistic intergroup relations involve aggression against outgroups with which collective narcissistic perceive as threatening. Collective narcissism helps to explain unreasonable manifestations of retaliation between groups. A narcissistic group is more sensitive to perceived criticism exhibited by outgroups, and is therefore more likely to retaliate. Collective narcissism is also related to negativity between groups who share a history of distressing experiences. The members of a narcissistic ingroup are likely to assume threats or negativity towards their ingroup where threats or negativity were not necessarily implied or exhibited. It is thought that this heightened sensitivity to negative feelings towards the ingroup is a result of underlying doubts about the greatness of the ingroup held by its members.

Similar to other elements of collective narcissism, intergroup aggression related to collective narcissism draws parallels with its individually narcissistic counterparts. An individual narcissist might react aggressively in the presence of humiliation, irritation, or anything threatening to his self-image. Likewise, a collective narcissist, or a collectively narcissistic group might react aggressively when the image of the group is in jeopardy, or when the group is collectively humiliated.

A study conducted among 6 to 9 year-olds by Judith Griffiths indicated that ingroups and outgroups among these children functioned relatively identical to other known collectively narcissistic groups in terms of intergroup aggression. The study noted that children generally had a significantly higher opinion of their ingroup than of surrounding outgroups, and that such ingroups indirectly or directly exhibited aggression on surrounding outgroups.

Ethnocentrism

Collective narcissism and ethnocentrism are closely related; they can be positively correlated and often shown to be coexistent, but they are independent in that either can exist without the presence of the other. In a study conducted by PhD Boris Bizumic, some ethnocentrism was shown to be an expression of group-level narcissism. It was noted, however, that not all manifestations of ethnocentrism are narcissistically based, and conversely, not all cases of group-level narcissism are by any means ethnocentric.

It is suggested that ethnocentrism, when pertaining to discrimination or aggression based on the self-love of one's group, or in other words, based on exclusion from one's self-perceived superior group is an expression of collective narcissism. In this sense, it might be said the collective and group narcissism overlap with ethnocentrism depending on given definitions, and the breadth of their acceptance.

In the world

In general, collective narcissism is most strongly manifested in groups that are "self-relevant", like religions, nationality, or ethnicity. As discussed earlier, phenomena such as national identity (nationality) and Nazi Germany (ethnicity and nationality) are manifestations of collective narcissism among groups that critically define the people who belong to them.

In addition to this, the collective narcissism that a group may already possess is likely to be exacerbated during conflict and aggression. And in terms of cultural effects, cultures that place an emphasis on the individual are apparently more likely to see manifestations of perceived individual greatness projected onto social ingroups existing within that culture. Also, and finally, narcissistic groups are not restricted to any one homogenous composition of collective or individually collective or individual narcissists. A quote from Hitler almost ideally sums the actual nature of collective narcissism as it is realistically manifested, and might be found reminiscent of almost every idea presented here: "My group is better and more important than other groups, but still is not worthy of me".

The Genius Neuroscientist Who Might Hold the Key to True AI

Author:  Shaun Raviv
Original linkhttps://www.wired.com/story/karl-friston-free-energy-principle-artificial-intelligence/



Karl Friston’s free energy principle might be the most all-encompassing idea since Charles Darwin’s theory of natural selection. But to understand it, you need to peer inside the mind of Friston himself.
Kate Peters


When King George III of England began to show signs of acute mania toward the end of his reign, rumors about the royal madness multiplied quickly in the public mind. One legend had it that George tried to shake hands with a tree, believing it to be the King of Prussia. Another described how he was whisked away to a house on Queen Square, in the Bloomsbury district of London, to receive treatment among his subjects. The tale goes on that George’s wife, Queen Charlotte, hired out the cellar of a local pub to stock provisions for the king’s meals while he stayed under his doctor’s care.
 
More than two centuries later, this story about Queen Square is still popular in London guidebooks. And whether or not it’s true, the neighborhood has evolved over the years as if to conform to it. A metal statue of Charlotte stands over the northern end of the square; the corner pub is called the Queen’s Larder; and the square’s quiet rectangular garden is now all but surrounded by people who work on brains and people whose brains need work. The National Hospital for Neurology and Neurosurgery—where a modern-day royal might well seek treatment—dominates one corner of Queen Square, and the world-renowned neuroscience research facilities of University College London round out its perimeter. During a week of perfect weather last July, dozens of neurological patients and their families passed silent time on wooden benches at the outer edges of the grass.

On a typical Monday, Karl Friston arrives on Queen Square at 12:25 pm and smokes a cigarette in the garden by the statue of Queen Charlotte. A slightly bent, solitary figure with thick gray hair, Friston is the scientific director of University College London’s storied Functional Imaging Laboratory, known to everyone who works there as the FIL. After finishing his cigarette, Friston walks to the western side of the square, enters a brick and limestone building, and heads to a seminar room on the fourth floor, where anywhere from two to two dozen people might be facing a blank white wall waiting for him. Friston likes to arrive five minutes late, so everyone else is already there.

His greeting to the group is liable to be his first substantial utterance of the day, as Friston prefers not to speak with other human beings before noon. (At home, he will have conversed with his wife and three sons via an agreed-upon series of smiles and grunts.) He also rarely meets people one-on-one. Instead, he prefers to hold open meetings like this one, where students, postdocs, and members of the public who desire Friston’s expertise—a category of person that has become almost comically broad in recent years—can seek his knowledge. “He believes that if one person has an idea or a question or project going on, the best way to learn about it is for the whole group to come together, hear the person, and then everybody gets a chance to ask questions and discuss. And so one person’s learning becomes everybody’s learning,” says David Benrimoh, a psychiatry resident at McGill University who studied under Friston for a year. “It’s very unique. As many things are with Karl.”

The A.I. Issue

At the start of each Monday meeting, everyone goes around and states their questions at the outset. Friston walks in slow, deliberate circles as he listens, his glasses perched at the end of his nose, so that he is always lowering his head to see the person who is speaking. He then spends the next few hours answering the questions in turn. “A Victorian gentleman, with Victorian manners and tastes,” as one friend describes Friston, he responds to even the most confused questions with courtesy and rapid reformulation. The Q&A sessions—which I started calling “Ask Karl” meetings—are remarkable feats of endurance, memory, breadth of knowledge, and creative thinking. They often end when it is time for Friston to retreat to the minuscule metal balcony hanging off his office for another smoke.

Friston first became a heroic figure in academia for devising many of the most important tools that have made human brains legible to science. In 1990 he invented statistical parametric mapping, a computational technique that helps—as one neuroscientist put it—“squash and squish” brain images into a consistent shape so that researchers can do apples-to-apples comparisons of activity within different crania. Out of statistical parametric mapping came a corollary called voxel-­based morphometry, an imaging technique that was used in one famous study to show that the rear side of the hippocampus of London taxi drivers grew as they learned “the knowledge.” To earn a London taxi license, drivers must memorize 320 routes and many landmarks within 6 miles of Charing Cross. The grueling process includes a written test as well as a series of one-on-one meetings with an examiner.

A study published in Science in 2011 used yet a third brain-imaging-analysis software invented by Friston—dynamic causal modeling—to determine if people with severe brain damage were minimally conscious or simply vegetative.

When Friston was inducted into the Royal Society of Fellows in 2006, the academy described his impact on studies of the brain as “revolutionary” and said that more than 90 percent of papers published in brain imaging used his methods. Two years ago, the Allen Institute for Artificial Intelligence, a research outfit led by AI pioneer Oren Etzioni, calculated that Friston is the world’s most frequently cited neuroscientist. He has an h-­index—a metric used to measure the impact of a researcher’s publications—nearly twice the size of Albert Einstein’s. Last year Clarivate Analytics, which over more than two decades has successfully predicted 46 Nobel Prize winners in the sciences, ranked Friston among the three most likely winners in the physiology or medicine category.

What’s remarkable, however, is that few of the researchers who make the pilgrimage to see Friston these days have come to talk about brain imaging at all. Over a 10-day period this summer, Friston advised an astrophysicist, several philosophers, a computer engineer working on a more personable competitor to the Amazon Echo, the head of artificial intelligence for one of the world’s largest insurance companies, a neuroscientist seeking to build better hearing aids, and a psychiatrist with a startup that applies machine learning to help treat depression. And most of them had come because they were desperate to understand something else entirely.

For the past decade or so, Friston has devoted much of his time and effort to developing an idea he calls the free energy principle. (Friston refers to his neuroimaging research as a day job, the way a jazz musician might refer to his shift at the local public library.) With this idea, Friston believes he has identified nothing less than the organizing principle of all life, and all intelligence as well. “If you are alive,” he sets out to answer, “what sorts of behaviors must you show?”

First the bad news: The free energy principle is maddeningly difficult to understand. So difficult, in fact, that entire rooms of very, very smart people have tried and failed to grasp it. A Twitter account with 3,000 followers exists simply to mock its opacity, and nearly every person I spoke with about it, including researchers whose work depends on it, told me they didn’t fully comprehend it. The account is called @FarlKriston. Sample tweet: “Life is an inevitable & emergent property of any (ergodic) random dynamical system that possesses a Markov blanket. Don’t leave with out it!”

But often those same people hastened to add that the free energy principle, at its heart, tells a simple story and solves a basic puzzle. The second law of thermodynamics tells us that the universe tends toward entropy, toward dissolution; but living things fiercely resist it. We wake up every morning nearly the same person we were the day before, with clear separations between our cells and organs, and between us and the world without. How? Friston’s free energy principle says that all life, at every scale of organization—from single cells to the human brain, with its billions of neurons—is driven by the same universal imperative, which can be reduced to a mathematical function. To be alive, he says, is to act in ways that reduce the gulf between your expectations and your sensory inputs. Or, in Fristonian terms, it is to minimize free energy.

To get a sense of the potential implications of this theory, all you have to do is look at the array of people who darken the FIL’s doorstep on Monday mornings. Some are here because they want to use the free energy principle to unify theories of the mind, provide a new foundation for biology, and explain life as we know it. Others hope the free energy principle will finally ground psychiatry in a functional understanding of the brain. And still others come because they want to use Friston’s ideas to break through the roadblocks in artificial intelligence research. But they all have one reason in common for being here, which is that the only person who truly understands Karl Friston’s free energy principle may be Karl Friston himself.

Friston’s office. A friend describes him as “a Victorian gentleman, with Victorian manners and tastes.” Kate Peters

Friston isn't just one of the most influential scholars in his field; he’s also among the most prolific in any discipline. He is 59 years old, works every night and weekend, and has published more than 1,000 academic papers since the turn of the millennium. In 2017 alone, he was a lead or coauthor of 85 publications—which amounts to approximately one every four days.

A 2018 article in Nature analyzed the phenomenon of “hyperprolific” scholars, which the authors defined as anyone with more than 72 publications in a year.

But if you ask him, this output isn’t just the fruit of an ambitious work ethic; it’s a mark of his tendency toward a kind of rigorous escapism.

Friston draws a carefully regulated boundary around his inner life, guarding against intrusions, many of which seem to consist of “worrying about other people.” He prefers being onstage, with other people at a comfortable distance, to being in private conversations. He does not have a mobile phone. He always wears navy-blue suits, which he buys two at a time at a closeout shop. He finds disruptions to his weekly routine on Queen Square “rather nerve-racking” and so tends to avoid other human beings at, say, international conferences. He does not enjoy advocating for his own ideas.

At the same time, Friston is exceptionally lucid and forthcoming about what drives him as a scholar. He finds it incredibly soothing—not unlike disappearing for a smoke—to lose himself in a difficult problem that takes weeks to resolve. And he has written eloquently about his own obsession, dating back to childhood, with finding ways to integrate, unify, and make simple the apparent noise of the world.

Friston traces his path to the free energy principle back to a hot summer day when he was 8 years old. He and his family were living in the walled English city of Chester, near Liverpool, and his mother had told him to go play in the garden. He turned over an old log and spotted several wood lice—small bugs with armadillo-shaped exoskeletons—moving about, he initially assumed, in a frantic search for shelter and darkness. After staring at them for half an hour, he deduced that they were not actually seeking the shade. “That was an illusion,” Friston says. “A fantasy that I brought to the table.”

He realized that the movement of the wood lice had no larger purpose, at least not in the sense that a human has a purpose when getting in a car to run an errand. The creatures’ movement was random; they simply moved faster in the warmth of the sun.

Young Friston was probably right. Many species of wood lice will dry out in direct sunlight, and some respond to a rise in temperature with kinesis, an increased level of random movement.

Friston calls this his first scientific insight, a moment when “all these contrived, anthropomorphized explanations of purpose and survival and the like all seemed to just peel away,” he says. “And the thing you were observing just was. In the sense that it could be no other way.”

Friston’s father was a civil engineer who worked on bridges all around England, and his family moved around with him. In just his first decade, the young Friston attended six different schools. His teachers often didn’t know what to do with him, and he drew most of his fragile self-esteem from solitary problem solving. At age 10 he designed a self-righting robot that could, in theory, traverse uneven ground while carrying a glass of water, using self-correcting feedback actuators and mercury levels. At school, a psychologist was brought in to ask him how he came up with it. “You’re very intelligent, Karl,” Friston’s mother reassured him, not for the last time. “Don’t let them tell you you’re not.” He says he didn’t believe her.

When Friston was in his mid-teens, he had another wood-lice moment. He had just come up to his bedroom from watching TV and noticed the cherry trees in bloom outside the window. He suddenly became possessed by a thought that has never let go of him since. “There must be a way of understanding everything by starting from nothing,” he thought. “If I’m only allowed to start off with one point in the entire universe, can I derive everything else I need from that?” He stayed there on his bed for hours, making his first attempt. “I failed completely, obviously,” he says.

Toward the end of secondary school, Friston and his classmates were the subjects of an early experiment in computer-­assisted advising. They were asked a series of questions, and their answers were punched into cards and run through a machine to extrapolate the perfect career choice. Friston had described how he enjoyed electronics design and being alone in nature, so the computer suggested he become a television antenna installer. That didn’t seem right, so he visited a school career counselor and said he’d like to study the brain in the context of mathematics and physics. The counselor told Friston he should become a psychiatrist, which meant, to Friston’s horror, that he had to study medicine.

Both Friston and the counselor had confused psychiatry with psychology, which is what he probably ought to have pursued as a future researcher. But it turned out to be a fortunate error, as it put Friston on a path toward studying both the mind and body,5 and toward one of the most formative experiences of his life—one that got Friston out of his own head.

Friston found time for other pursuits as well. At age 19, he spent an entire school vacation trying to squeeze all of physics on one page. He failed but did manage to fit all of quantum mechanics.

After completing his medical studies, Friston moved to Oxford and spent two years as a resident trainee at a Victorian-era hospital called Littlemore. Founded under the 1845 Lunacy Act, Littlemore had originally been instituted to help transfer all “pauper lunatics” from workhouses to hospitals. By the mid-1980s, when Friston arrived, it was one of the last of the old asylums on the outskirts of England’s cities.

Friston was assigned a group of 32 chronic schizophrenic patients, the worst-off residents of Littlemore, for whom treatment mostly meant containment. For Friston, who recalls his former patients with evident nostalgia, it was an introduction to the way that connections in the brain were easily broken. “It was a beautiful place to work,” he says. “This little community of intense and florid psychopathology.”

Twice a week he led 90-minute group therapy sessions in which the patients explored their ailments together, reminiscent of the Ask Karl meetings today. The group included colorful characters who still inspire Friston’s thinking more than 30 years later. There was Hillary, who looked like she could play the senior cook on Downton Abbey but who, before coming to Littlemore, had decapitated her neighbor with a kitchen knife, convinced he had become an evil, human-sized crow.

The names of Friston’s patients at Littlemore have been changed in this story.

There was Ernest, who had a penchant for pastel Marks & Spencer cardigans and matching plimsoll shoes, and who was “as rampant and incorrigible a pedophile as you could ever imagine,” Friston says.

And then there was Robert, an articulate young man who might have been a university student had he not suffered severe schizophrenia. Robert ruminated obsessively about, of all things, angel shit; he pondered whether the stuff was a blessing or a curse and whether it was ever visible to the eye, and he seemed perplexed that these questions had not occurred to others. To Friston, the very concept of angel shit was a miracle. It spoke to the ability of people with schizophrenia to assemble concepts that someone with a more regularly functioning brain couldn’t easily access. “It’s extremely difficult to come up with something like angel shit,” Friston says with something like admiration. “I couldn’t do it.”

After Littlemore, Friston spent much of the early 1990s using a relatively new technology—PET scans—to try to understand what was going on inside the brains of people with schizophrenia. He invented statistical parametric mapping along the way. Unusually for the time, Friston was adamant that the technique should be freely shared rather than patented and commercialized, which largely explains how it became so widespread. Friston would fly across the world—to the National Institutes of Health in Bethesda, Maryland, for example—to give it to other researchers. “It was me, literally, with a quarter of biometric tape, getting on an airplane, taking it over there, downloading it, spending a day getting it to work, teaching somebody how to use it, then going home for a rest,” Friston says. “This is how open source software worked in those days.”

Friston came to Queen Square in 1994, and for a few years his office at the FIL sat just a few doors down from the Gatsby Computational Neuroscience Unit. The Gatsby—where researchers study theories of perception and learning in both living and machine systems—was then run by its founder, the cognitive psychologist and computer scientist Geoffrey Hinton. While the FIL was establishing itself as one of the premier labs for neuroimaging, the Gatsby was becoming a training ground for neuroscientists interested in applying mathematical models to the nervous system.

Friston, like many others, became enthralled by Hinton’s “childlike enthusiasm” for the most unchildlike of statistical models, and the two men became friends.

At the time, Hinton was living in a particularly noisy building in Camden. The neighbors’ water pipes were so loud that he built a soundproof box in a basement bedroom out of rubber and ¾-inch drywall where he and his wife could sleep.

Over time, Hinton convinced Friston that the best way to think of the brain was as a Bayesian probability machine. The idea, which goes back to the 19th century and the work of Hermann von Helmholtz, is that brains compute and perceive in a probabilistic manner, constantly making predictions and adjusting beliefs based on what the senses contribute. According to the most popular modern Bayesian account, the brain is an “inference engine” that seeks to minimize “prediction error.”

In 2001, Hinton left London for the University of Toronto, where he became one of the most important figures in artificial intelligence, laying the groundwork8 for much of today’s research in deep learning.

In 2012, Hinton won the ImageNet Challenge, a competition to identify objects in a 15-million-image database built by Fei-Fei Li. ImageNet helped propel neural networks—and Hinton—to the forefront of AI.

Before Hinton left, however, Friston visited his friend at the Gatsby one last time. Hinton described a new technique he’d devised to allow computer programs to emulate human decisionmaking more efficiently—a process for integrating the input of many different probabilistic models, now known in machine learning as a “product of experts.”

The meeting left Friston’s head spinning. Inspired by Hinton’s ideas, and in a spirit of intellectual reciprocity, Friston sent Hinton a set of notes about an idea he had for connecting several seemingly “unrelated anatomical, physiological, and psychophysical attributes of the brain.” Friston published those notes in 2005—the first of many dozens of papers he would go on to write about the free energy principle.


The Markov blanket in Karl Friston’s office—“keeping your internal states warm since 1856.”
Kate Peters

Even Friston has a hard time deciding where to start when he describes the free energy principle. He often sends people to its Wikipedia page. But for my part, it seems apt to begin with the blanket draped over the futon in Friston’s office.

It’s a white fleece throw, custom-printed with a black-and-white portrait of a stern, bearded Russian mathematician named Andrei Andreyevich Markov, who died in 1922. The blanket is a gag gift from Friston’s son, a plush, polyester inside joke about an idea that has become central to the free energy principle. Markov is the eponym of a concept called a Markov blanket, which in machine learning is essentially a shield that separates one set of variables from others in a layered, hierarchical system. The psychologist Christopher Frith—who has an h-index on par with Friston’s—once described a Markov blanket as “a cognitive version of a cell membrane, shielding states inside the blanket from states outside.”

In Friston’s mind, the universe is made up of Markov blankets inside of Markov blankets. Each of us has a Markov blanket that keeps us apart from what is not us. And within us are blankets separating organs, which contain blankets separating cells, which contain blankets separating their organelles. The blankets define how biological things exist over time and behave distinctly from one another. Without them, we’re just hot gas dissipating into the ether.

“That’s the Markov blanket you’ve read about. This is it. You can touch it,” Friston said dryly when I first saw the throw in his office. I couldn’t help myself; I did briefly reach out to feel it under my fingers. Ever since I first read about Markov blankets, I’d seen them everywhere. Markov blankets around a leaf and a tree and a mosquito. In London, I saw them around the postdocs at the FIL, around the black-clad protesters at an antifascist rally, and around the people living in boats in the canals. Invisible cloaks around everyone, and underneath each one a different living system that minimizes its own free energy.

The concept of free energy itself comes from physics, which means it’s difficult to explain precisely without wading into mathematical formulas. In a sense that’s what makes it powerful: It isn’t a merely rhetorical concept. It’s a measurable quantity that can be modeled, using much the same math that Friston has used to interpret brain images to such world-­changing effect. But if you translate the concept from math into English, here’s roughly what you get: Free energy is the difference between the states you expect to be in and the states your sensors tell you that you are in. Or, to put it another way, when you are minimizing free energy, you are minimizing surprise.

According to Friston, any biological system that resists a tendency to disorder and dissolution will adhere to the free energy principle—whether it’s a protozoan or a pro basketball team.

In 2013, Friston ran a model that simulated a primordial soup full of floating molecules. He programmed it to obey both basic physics and the free energy principle. The model generated results that looked like organized life.

A single-celled organism has the same imperative to reduce surprise that a brain does.

The only difference is that, as self-organizing biological systems go, the human brain is inordinately complex: It soaks in information from billions of sense receptors, and it needs to organize that information efficiently into an accurate model of the world. “It’s literally a fantastic organ in the sense that it generates hypotheses or fantasies that are appropriate for trying to explain these myriad patterns, this flux of sensory information that it is in receipt of,” Friston says. In seeking to predict what the next wave of sensations is going to tell it—and the next, and the next—the brain is constantly making inferences and updating its beliefs based on what the senses relay back, and trying to minimize prediction-error signals.

So far, as you might have noticed, this sounds a lot like the Bayesian idea of the brain as an “inference engine” that Hinton told Friston about in the 1990s. And indeed, Friston regards the Bayesian model as a foundation of the free energy principle (“free energy” is even a rough synonym for “prediction error”). But the limitation of the Bayesian model, for Friston, is that it only accounts for the interaction between beliefs and perceptions; it has nothing to say about the body or action. It can’t get you out of your chair.

This isn’t enough for Friston, who uses the term “active inference” to describe the way organisms minimize surprise while moving about the world. When the brain makes a prediction that isn’t immediately borne out by what the senses relay back, Friston believes, it can minimize free energy in one of two ways: It can revise its prediction—absorb the surprise, concede the error, update its model of the world—or it can act to make the prediction true. If I infer that I am touching my nose with my left index finger, but my proprioceptors tell me my arm is hanging at my side, I can minimize my brain’s raging prediction-error signals by raising that arm up and pressing a digit to the middle of my face.

And in fact, this is how the free energy principle accounts for everything we do: perception, action, planning, problem solving. When I get into the car to run an errand, I am minimizing free energy by confirming my hypothesis—my fantasy—through action.

For Friston, folding action and movement into the equation is immensely important. Even perception itself, he says, is “enslaved by action”: To gather information, the eye darts, the diaphragm draws air into the nose, the fingers generate friction against a surface. And all of this fine motor movement exists on a continuum with bigger plans, explorations, and actions.

Friston’s term for this kind of exploration is “epistemic foraging.” He is notorious among his colleagues for his coinages, known as Fristonese.

“We sample the world,” Friston writes, “to ensure our predictions become a self-fulfilling prophecy.”

So what happens when our prophecies are not self-fulfilling? What does it look like for a system to be overwhelmed by surprise? The free energy principle, it turns out, isn’t just a unified theory of action, perception, and planning; it’s also a theory of mental illness. When the brain assigns too little or too much weight to evidence pouring in from the senses, trouble occurs. Someone with schizophrenia, for example, may fail to update their model of the world to account for sensory input from the eyes. Where one person might see a friendly neighbor, Hillary might see a giant, evil crow. “If you think about psychiatric conditions, and indeed most neurological conditions, they are just broken beliefs or false inference—hallucinations and delusions,” Friston says.

Over the past few years, Friston and a few other scientists have used the free energy principle to help explain anxiety, depression, and psychosis, along with certain symptoms of autism, Parkinson’s disease, and psychopathy. In many cases, scientists already know—thanks to Friston’s neuroimaging methods—which regions of the brain tend to malfunction in different disorders and which signals tend to be disrupted. But that alone isn’t enough to go on. “It’s not sufficient to understand which synapses, which brain connections, are working improperly,” he says. “You need to have a calculus that talks about beliefs.”

So: The free energy principle offers a unifying explanation for how the mind works and a unifying explanation for how the mind malfunctions. It stands to reason, then, that it might also put us on a path toward building a mind from scratch.

A few years ago, a team of British researchers decided to revisit the facts of King George III’s madness with a new analytic tool. They loaded some 500 letters written by the king into a machine-learning engine and laboriously trained the system to recognize various textual features: word repetition, sentence length, syntactical complexity, and the like. By the end of the training process, the system was able to predict whether a royal missive had been written during a period of mania or during a period of sanity.

This kind of pattern-matching technology—which is roughly similar to the techniques that have taught machines to recognize faces, images of cats, and speech patterns—has driven huge advances in computing over the past several years. But it requires a lot of up-front data and human supervision, and it can be brittle. Another approach to AI, called reinforcement learning, has shown incredible success at winning games: Go, chess, Atari’s Breakout. Reinforcement learning doesn’t require humans to label lots of training data; it just requires telling a neural network to seek a certain reward, often victory in a game. The neural network learns by playing the game over and over, optimizing for whatever moves might get it to the final screen, the way a dog might learn to perform certain tasks for a treat.

But reinforcement learning, too, has pretty major limitations. In the real world, most situations are not organized around a single, narrowly defined goal. (Sometimes you have to stop playing Breakout to go to the bathroom, put out a fire, or talk to your boss.) And most environments aren’t as stable and rule-bound as a game is. The conceit behind neural networks is that they are supposed to think the way we do; but reinforcement learning doesn’t really get us there.

To Friston and his enthusiasts, this failure makes complete sense. After all, according to the free energy principle, the fundamental drive of human thought isn’t to seek some arbitrary external reward. It’s to minimize prediction error. Clearly, neural networks ought to do the same. It helps that the Bayesian formulas behind the free energy principle—the ones that are so difficult to translate into English—are already written in the native language of machine learning.

Julie Pitt, head of machine-learning infrastructure at Netflix, discovered Friston and the free energy principle in 2014, and it transformed her thinking. (Pitt’s Twitter bio reads, “I infer my own actions by way of Active Inference.”) Outside of her work at Netflix, she’s been exploring applications of the principle in a side project called Order of Magnitude Labs. Pitt says that the beauty of the free energy model is that it allows an artificial agent to act in any environment, even one that’s new and unknown. Under the old reinforcement-learning model, you’d have to keep stipulating new rules and sub-rewards to get your agent to cope with a complex world. But a free energy agent always generates its own intrinsic reward: the minimization of surprise. And that reward, Pitt says, includes an imperative to go out and explore.

In late 2017, a group led by Rosalyn Moran, a neuroscientist and engineer at King’s College London, pitted two AI players against one another in a version of the 3D shooter game Doom. The goal was to compare an agent driven by active inference to one driven by reward-maximization.

The reward-based agent’s goal was to kill a monster inside the game, but the free-energy-driven agent only had to minimize surprise. The Fristonian agent started off slowly. But eventually it started to behave as if it had a model of the game, seeming to realize, for instance, that when the agent moved left the monster tended to move to the right.

After a while it became clear that, even in the toy environment of the game, the reward-­maximizing agent was “demonstrably less robust”; the free energy agent had learned its environment better. “It outperformed the reinforcement-­learning agent because it was exploring,” Moran says. In another simulation that pitted the free-­energy-minimizing agent against real human players, the story was similar. The Fristonian agent started slowly, actively exploring options—epistemically foraging, Friston would say—before quickly attaining humanlike performance.

Moran told me that active inference is starting to spread into more mainstream deep-­learning research, albeit slowly. Some of Friston’s students have gone on to work at DeepMind and Google Brain, and one of them founded Huawei’s Artificial Intelligence Theory lab. “It’s moving out of Queen Square,” Moran says. But it’s still not nearly as common as reinforcement learning, which even undergraduates learn. “You don’t teach undergraduates the free energy principle—yet.”

The first time I asked Friston about the connection between the free energy principle and artificial intelligence, he predicted that within five to 10 years, most machine learning would incorporate free energy minimization. The second time, his response was droll. “Think about why it’s called active inference,” he said. His straight, sparkly white teeth showed through his smile as he waited for me to follow his wordplay. “Well, it’s AI,” Friston said. “So is active inference the new AI? Yes, it’s the acronym.” Not for the first time, a Fristonian joke had passed me by.

While I was in London, Friston gave a talk at a quantitative trading firm. About 60 baby-faced stock traders were in attendance, rounding out the end of their workday. Friston described how the free energy principle could model curiosity in artificial agents. About 15 minutes in, he asked his listeners to raise a hand if they understood what he was saying. He counted only three hands, so he reversed the question: “Can you put your hand up if that was complete nonsense and you don’t know what I was talking about?” This time, a lot of people raised their hands, and I got the feeling that the rest were being polite. With 45 minutes left, Friston turned to the organizer of the talk and looked at him as if to say, What the hell? The manager stammered a bit before saying, “Everybody here’s smart.” Friston graciously agreed and finished his presentation.

The next morning, I asked Friston if he thought the talk went well, considering that few of those bright young minds seemed to understand him. “There is going to be a substantial proportion of the audience who—it’s just not for them,” he said. “Sometimes they get upset because they’ve heard that it’s important and they don’t understand it. They think they have to think it’s rubbish and they leave. You get used to that.”

In 2010, Peter Freed, a psychiatrist at Columbia University, gathered together 15 brain researchers to discuss one of Friston’s papers. Freed described what happened in the journal Neuropsychoanalysis: “There was a lot of mathematical knowledge in the room: three statisticians, two physicists, a physical chemist, a nuclear physicist, and a large group of neuroimagers—but apparently we didn’t have what it took. I met with a Princeton physicist, a Stanford neurophysiologist, a Cold Springs Harbor neurobiologist to discuss the paper. Again blanks, one and all: too many equations, too many assumptions, too many moving parts, too global a theory, no opportunity for questions—and so people gave up.”

But for all the people who are exasperated by Friston’s impenetrability, there are nearly as many who feel he has unlocked something huge, an idea every bit as expansive as Darwin’s theory of natural selection. When the Canadian philosopher Maxwell Ramstead first read Friston’s work in 2014, he had already been trying to find ways to connect complex living systems that exist at different scales—from cells to brains to individuals to cultures. In 2016 he met Friston, who told him that the same math that applies to cellular differentiation—the process by which generic cells become more specialized—can also be applied to cultural dynamics. “This was a life-changing conversation for me,” Ramstead says. “I almost had a nosebleed.”

“This is absolutely novel in history,” Ramstead told me as we sat on a bench in Queen Square, surrounded by patients and staff from the surrounding hospitals. Before Friston came along, “We were kind of condemned to forever wander in this multidisciplinary space without a common currency,” he continued. “The free energy principle gives you that currency.”

In 2017, Ramstead and Friston coauthored a paper, with Paul Badcock of the University of Melbourne, in which they described all life in terms of Markov blankets. Just as a cell is a Markov-blanketed system that minimizes free energy in order to exist, so are tribes and religions and species.

After the publication of Ramstead’s paper, Micah Allen, a cognitive neuroscientist then at the FIL, wrote that the free energy principle had evolved into a real-life version of Isaac Asimov’s psychohistory, a fictional system that reduced all of psychology, history, and physics down to a statistical science.

In Foundation, published in 1951, one of Asimov’s characters defines psychohistory as “that branch of mathematics which deals with the reactions of human conglomerates to fixed social and economic stimuli.”

And it’s true that the free energy principle does seem to have expanded to the point of being, if not a theory of everything, then nearly so. (Friston told me that cancer and tumors might be instances of false inference, when cells become deluded.) As Allen asked: Does a theory that explains everything run the risk of explaining nothing?

On the last day of my trip, I visited Friston in the town of Rickmansworth, where he lives in a house filled with taxidermied animals that his wife prepares as a hobby.

On a recent Saturday, a man came to the door asking if Friston’s wife was home. When Friston said yes, the man said, “Good, because I got a dead cat here.” He wanted it stuffed.

As it happens, Rickmansworth appears on the first page of The Hitchhiker’s Guide to the Galaxy; it’s the town where “a girl sitting on her own in a small café” suddenly discovers the secret to making the world “a good and happy place.” But fate intervenes. “Before she could get to a phone to tell anyone about it, a terrible stupid catastrophe occurred, and the idea was lost forever.”

It’s unclear whether the free energy principle is the secret to making the world a good and happy place, as some of its believers almost seem to think it might be. Friston himself tended to take a more measured tone as our talks went on, suggesting only that active inference and its corollaries were quite promising. Several times he conceded that he might just be “talking rubbish.” During the last group meeting I attended at the FIL, he told those in attendance that the free energy principle is an “as if” concept—it does not require that biological things minimize free energy in order to exist; it is merely sufficient as an explanation for biotic self-organization.

Friston’s mother died a few years ago, but lately he has been thinking back to her frequent reassurances during his childhood: You’re very intelligent, Karl. “I never quite believed her,” he says. “And yet now I have found myself suddenly being seduced by her argument. Now I do believe I’m actually quite bright.” But this newfound self-esteem, he says, has also led him to examine his own egocentricity.

Friston says his work has two primary motivations. Sure, it would be nice to see the free energy principle lead to true artificial consciousness someday, he says, but that’s not one of his top priorities. Rather, his first big desire is to advance schizophrenia research, to help repair the brains of patients like the ones he knew at the old asylum. And his second main motivation, he says, is “much more selfish.” It goes back to that evening in his bedroom, as a teenager, looking at the cherry blossoms, wondering, “Can I sort it all out in the simplest way possible?

“And that is a very self-indulgent thing. It has no altruistic clinical compassion behind it. It is just the selfish desire to try and understand things as completely and as rigorously and as simply as possible,” he says. “I often reflect on the jokes that people make about me—sometimes maliciously, sometimes very amusingly—that I can’t communicate. And I think: I didn’t write it for you. I wrote it for me.”

Friston told me he occasionally misses the last train home to Rickmansworth, lost in one of those problems that he drills into for weeks. So he’ll sleep in his office, curled on the futon under his Markov blanket, safe and securely separated from the external world.



Shaun Raviv (@ShaunRaviv) is a writer living in Atlanta, Georgia.
This article appears in the December issue. Subscribe now.
Let us know what you think about this article. Submit a letter to the editor at mail@wired.com.

Gene

From Wikipedia, the free encyclopedia https://en.wikipedia.org/wiki/Gene Chromosome ...