Denmark

Introduction
Many domains of knowledge have grappled with the arrival of hypercompetent artificial intelligence. We might expect an efflorescence of AI-produced ideas. On the other hand, human-trained AI might simply reproduce or reinforce existing orthodoxies. The strategic impact of AlphaGo post-2016 on the game of Go serves as a data-rich case study.
&lt;br&gt;Methods
Using a large database of top-level human and AI play (62,740 opening sequences, 1956-2020), I study the diversity of opening sequences using multilevel regression models of recent social and individual covariates pre &amp; post-AlphaGo.&lt;br&gt;
Results
The post-AI era has indeed seen an explosion of move diversity (Shannon entropy on opening sequences), reversing decades of decline. Since the 1990s, social trends have come to strongly predict individual first-move openings (b=3.77, psign&lt;0.001), and, while pre-internet Go was characterized by a low correlation between strategy popularity and efficacy (R2=0.11), the post-AI era now sees a near-perfect correlation (R2=0.87).
&lt;br&gt;Conclusions
Human Go strategies are more diverse, more social, and apparently more efficacious under the influence of hypercompetent AI advisors.

Cultural Evolution Society 2022

Strategic Social Leaning in the Era of Hypercompetent Artificial Intelligence: A Large-Scale, Longitudinal Case Study using the Game of Go

strategic social learning; cultural evolution; artificial intelligence; games; longitudinal databases

AAMAS (International Conference on Autonomous Agents and Multiagent Systems) is the largest and most influential conference in the area of agents and multiagent systems. The aim of the conference is to bring together researchers and practitioners in all areas of agent technology and to provide a single, high-profile, internationally renowned forum for research in the theory and practice of autonomous agents and multiagent systems. AAMAS is the flagship conference of the non-profit International Foundation for Autonomous Agents and Multiagent Systems (IFAAMAS).

The AAMAS conference series was initiated in 2002 in Bologna, Italy as a joint event comprising the 6th International Conference on Autonomous Agents (AA), the 5th International Conference on Multiagent Systems (ICMAS), and the 9th International Workshop on Agent Theories, Architectures, and Languages (ATAL).

Subsequent AAMAS conferences have been held in Melbourne, Australia (July 2003), New York City, NY, USA (July 2004), Utrecht, The Netherlands (July 2005), Hakodate, Japan (May 2006), Honolulu, Hawaii, USA (May 2007), Estoril, Portugal (May 2008), Budapest, Hungary (May 2009), Toronto, Canada (May 2010), Taipei, Taiwan (May 2011), Valencia, Spain (June 2012), Minnesota, USA (May 2013), Paris, France (May 2014), Istanbul, Turkey (May 2015), Singapore (May 2016), São Paulo (2017) and Stockholm, Sweden (2018), Montreal (May 2019), Auckland (May 2020, Virtual), London (May 2021, Virtual).

A ticket is required to attend this event, please register using the link below:

https://aamas2022-conference.auckland.ac.nz/attending/registration/

Registration Is Required

AAMAS 2022

AAMAS (International Conference on Autonomous Agents and Multiagent Systems) is the largest and most influential conference in the area of agents and multiagent systems. The aim of the conference is to bring together researchers and practitioners in all areas of agent technology and to provide a single, high-profile, internationally renowned forum for research in the theory and practice of autonomous agents and multiagent systems. AAMAS is the flagship conference of the non-profit International Foundation for Autonomous Agents and Multiagent Systems (IFAAMAS).

Introduction
Many domains of knowledge have grappled with the arrival of hypercompetent artificial intelligence. We might expect an efflorescence of AI-produced ideas. On the other hand, human-trained AI might simply reproduce or reinforce existing orthodoxies. The strategic impact of AlphaGo post-2016 on the game of Go serves as a data-rich case study.
 Methods
Using a large database of top-level human and AI play (62,740 opening sequences, 1956-2020), I study the diversity of opening sequences using multilevel regression models of recent social and individual covariates pre & post-AlphaGo. 
Results
The post-AI era has indeed seen an explosion of move diversity (Shannon entropy on opening sequences), reversing decades of decline. Since the 1990s, social trends have come to strongly predict individual first-move openings (b=3.77, psign<0.001), and, while pre-internet Go was characterized by a low correlation between strategy popularity and efficacy (R2=0.11), the post-AI era now sees a near-perfect correlation (R2=0.87).
 Conclusions
Human Go strategies are more diverse, more social, and apparently more efficacious under the influence of hypercompetent AI advisors.

technical paper

### Welcome to the fourth Cultural Evolution Society Conference, CES2022! 

As last year, Underline will support the online component of the conference. The ongoing COVID-19 pandemic is causing a great deal of uncertainty for all of us. Therefore, we seek to offer a hybrid format that will hopefully allow some folks to travel to Aarhus, others to meet up locally or regionally in CES hub-lets, and other yet to participate online. The field of cultural evolution is growing in size and scope; its theories, methods and applications - both academic and applied - are getting ever more robust. Do join us in what promises to be an exciting exchange of ideas and a celebration of our field.

##### ***For SoMe interaction and live tweeting please tag @CulturalEvolSoc and use #CESconf2022***


#### The CES 2022 local organising committee
* Luseadra Joy McKerracher
* Jesper Sørensen 
* Niels Nørkjær Johannsen 
* Benjamin Grant Purzycki 
* Mette Løvschal 
* Felix Riede
* Iza Romanowska 
* Karsten Olsen

Please click on the button bellow if you want to watch the videos

The field of cultural evolution is growing in size and scope; its theories, methods and applications - both academic and applied - are getting ever more robust. Do join us in what promises to be an exciting exchange of ideas and a celebration of our field.

 I propose a mathematical model of cultural evolution to explain why false theories and meaning systems or maladaptive beliefs may spread and become fixed in a population while assuming that their diffusion rates are lower than their true counterparts. By combining ideas from judgment aggregation and the Condorcet Jury theorem with cultural evolution dynamics, I introduce a new epistemic stage in the cultural evolution process where culturally transmitted theories not only spread “horizontally” (across individuals in the same generation) and “vertically” (from one generation to the next), but also inside individual brains (“personal reflection process”). This occurs when a brain adopts a maladaptive theory (not received from outside) by means of internal inference from other false premise/idea that was indeed received culturally. In this context, false theories may favor the assumption of other false theories after the individuals have been exposed to true information about their logical connection, and increase their presence in the population within the normal and adaptive logical working of our brains. Using a simple stylized model and subsequent simulations I identify the parametric conditions in which false theories become universally accepted, so that the model could be tested empirically.

The evolution of false theories

**Author(s):** Sobchuk, O; Youngblood, M; Morin, O Why do some artists achieve extreme popularity? A theory of first-mover advantage, developed in economics, may partly explain this. It suggests that companies that enter a new market niche early on become more successful. We hypothesize that the first-mover advantage can explain the popularity of songs. A musical genre is analogous to a niche, and early representatives of a genre may be considered first movers.

We use Spotify data: ~166,000 songs in 10 genres. To account for data biases, we do bootstrapping from a control mixed-genre dataset, built by sampling, for each genre, song corpora that are similar to the target genre: they have the exact same size and yearly distribution of releases but do not belong to any single genre. We have preregistered the prediction that the relationship between success (i.e. playcount) and time for the genre corpora should be more negative than for the samples taken from the mixed-genre corpora.

We find a more negative relationship between time and playcount for 8 genres, which supports the prediction; in the remaining 2 genres the difference is not significant.

Successful artists may owe their popularity to innovativeness in new artistic niches.

First-mover advantage in the evolution of music genres

**Author(s):** Turner, MA; Moya, C; Smaldino, PE; Jones, JH Introduction Social learning is essential to human survival–it is a critical adaptation for dealing with uncertainty. However, uncertainty takes many forms. We identified four theoretically important types of uncertainty: temporal environmental variability, ambiguous payoffs, decision set size, and effective lifespans. We used this framework to study conditions for the evolution of social learning in populations with

Different forms of uncertainty differently affect the evolution of social learning

 This paper will discuss the potential of using methods from digital humanities when studying the evolution of religions. The case at hand will be ongoing work on my doctoral thesis, in which I study separate developments in symbolic systems across the Indo-European religions. Digital methods, such as topic modelling, make it possible for scholars to look for patterns in large textual corpora. Combined with close philological readings, I believe this provides a more nuanced look into the symbolic system, religion, expressed in ancient texts. As an example, I will present my preliminary studies of the different layers of the Vedic Rigveda compared with the Old Iranian Gathas. In this study, digital methods will be an important tool in analysing patterns that unite and differentiate throughout the texts. When comparing the texts, these patterns may lead to not just reconstructions of earlier stages, but how these stages developed into the known religions. Thus, providing a better understanding of the shared Indo-Iranian heritage between the two texts, and the cultural evolutionary processes in the development of the individual symbolic systems expressed. As such, this paper will argue for the advantages of using digital methods when combining Indo-European studies and theories on religious evolution to view similarities, differences, and developments in the symbolic systems of large corpora.

Convenient collaborations: Using digital humanities in the study of the evolution of Indo-European religions

**Author(s):** Barrett, B; Kandler, A; Aplin, L Cultural evolution is usually framed in terms of transmission biases or behavioral transformation via cognitive biases, although there is a third dimension that is often ignored: learning through reward. This process of internal selection that happens when socially learned behaviors are used by individuals, has been hypothesized to play a critical role in the evolution of adaptiveness of cultural traits, as they are either maintained or abandoned over the course of repeated production. To date, this has remained a verbal argument. To test this argument, we construct an agent based model where acquisition and production are linked, as acquisition is conditioned on association, as well as behavioral frequencies of associates. We show that variation in the learning rules that govern production influence the diffusion of a novel behavior. We further show that these effects can affect the inferential power of two popular models in the field of animal culture: Network Based Diffusion Analysis and Experience Weighted Attraction models. We then allow for dynamic networks of populations, and show how the social process of population turnover interacts with reinforcement learning to drive the cultural evolution of efficiency. Different regimes of population turnover generate optimal, neutral or sub-optimal selection for a more efficient behavior, when compared against static populations. These two studies generate new predictions for how reinforcement learning can influence cultural outcomes, and highlight un-discussed consequences for current practices in the field. The formulation of our model also encourages more rigorous definitions for social learning strategies and biases, as they might target the process of acquisition or production, or both, yet have been discussed interchangeably to date. They also provide theoretical support for the hypothesis that naive learners, introduced through population turnover, play a crucial role in resampling payoff landscapes.

The interaction of learning rules and population turnover can drive cultural evolution

 Two important hypotheses compete to explain the human adaptation. The cognitive niche hypothesis proposes that gene-based evolution in the Pleistocene favored the evolution of information rich encapsulated cognitive modules adapted to the regularities of the Pleistocene environment. The cultural niche hypothesis argues that the hallmark of Pleistocene environments was extreme variability on millennial and submillennial time scales and that culture is an adaptation this. These hypotheses have very different implications for how the human mind/brain would work. Some neurobiologists for some time have argued that the cerebral cortex has to operate as an organ of phenotypic flexibility and that genes specify, in any detail, only the ancient emotional and vegetative parts of the brain. Recent neurobiological advances support this latter picture of cortical function. The ancient but rather generalized appetites and emotions provide the selective forces on the developmental time scale that also act as forces on the cultural evolutionary time scale. These advances tend to support the cultural niche hypothesis and are not consistent with the cognitive niche alternative.

The cognitive niche versus cultural niche hypotheses: The neurobiological evidence

**Author(s):** Chira, A; Kirby, K.; Epperlein, T.; Kuehr, F; Bräuer, J. Introduction: Dogs show an extraordinaire capacity to understand, communicate, and cooperate with us. However, what we know about the social-cognitive skills of dogs comes from observations and experiments performed in WEIRD societies. Around the world though, dogs are kept for various functions, they are perceived and treated differently, suggesting cross-cultural diversity in the strength of dog-human relationships, and perhaps departures from the ‘typical’ suite of skills dogs are thought to possess. Materials & Methods: Here we address these issues by collecting data on the function and perception of dogs in 124 globally distributed societies using the eHRAF cross-cultural database. We hypothesize that dog-human relationships will be closer in societies where dog functions require intense cooperation between dogs and humans, i.e., hunting and herding versus guarding, defense, and carry animals. We characterize close dog-human relationships as increased primary caregiving (or positive care), decreased negative treatment, and an increased extent to which dogs are viewed as friends or kin. Results & Discussion: The number of functions dogs perform associates with closer dog-human relationships. Further, the presence of herding (and to a lesser extent, guarding herds, and carry roles) increases the odds of positive care given to dogs, while using dogs for hunting is linked with higher odds of dogs being treated akin to a person. These findings support the idea of cross-cultural variation in dog-human relationships and show how the strength of our bond with dogs can be linked with the function dogs fill in societies. The results are a first step towards challenging the notion that all dogs are the same, and fuel questions about the universality of cognitive and social skills we commonly associate with our canine friends.

Beyond WEIRD dogs - Cultural Differences In Dog Human Interactions

**Author(s):** Peacey, S; Campbell, O.L.K.; Mace, R Introduction Cultures vary in whether they accuse more men or women of witchcraft (harming others through magic). We investigated why accusations target a specific sex in African Bantu and Bantoid societies. Accusations may be competitive, allowing accusers to access resources that the ‘witch’ could have used. For example, men may target male relatives when competing for wealth to marry; co-wives in polygynous marriages vying for their husband’s investment may accuse other wives.
Materials and Methods Data was coded from ethnographic records. We examined factors predicting whether ‘witches’ were male or female (N = 423) using multilevel logistic regression. 
Results Men commonly accused men: such allegations were frequently between nonrelatives and blood relatives, in disputes over inheritance or property. Accusations of women were often from affinal kin (mostly husbands and co-wives). 
Discussion/conclusion Associations between specific relationships and situations and the sex of accused witches suggest how ‘phenotypes’ may develop within societies. The results are broadly consistent with evolutionary patterns of conflict and competition.

Same-sex competition and sexual conflict expressed through witchcraft accusations

**Author(s):** Pichkar, Y; Creanza, N Introduction: Human culture is both a bearer of history and an active force in it. The history of cultures can resemble that of our genes, but the impact of cultural identity on long-term patterns of human interaction has not been extensively researched. We leverage increasingly available genetic and cultural databases to explore how culture has shaped populations in Great Britain. Approach: We used the Survey of English Dialects to produce a map of linguistic variation, a type of cultural variation. From the genetic data collected for the People of the British Isles project, we reconstruct previously found genetic clustering (Leslie et al. 2015). Using the genetic variation within and between clusters, we measured whether genetic variation coincides with dialect borders. Results: We find that these patterns are strongest in areas with larger cultural divides–near Scotland and Wales–but correlations also exist closer to the center of England. Conclusions: These patterns suggest that cultural boundaries have shaped people’s preferences for movement and marriage over the last millennium, and that cultural identity may contribute to establishing genetic differentiation in a region.

Cultural reinforcement of genetic boundaries in England

**Author(s):** Raymond C; Sánchez-Villagra MR; Rangel-de Lázaro G; Lynch, S; Ward S; Wilson LAB Intro: Intentional cranial modification (ICM), the application of materials to the cranium of infants to permanently alter the head shape, has a long history and is found throughout the world. We collated global records of ICM to trace the origin and dispersal of these practices. 
Methods: We collated published records for Central (n=271), South (n=53), and North America (n=54), the Caribbean (n=113), and Eurasia (n=252), recording geographic location (decimal coordinates), ICM type, biological sex, period, and date. The temporal sequence of ICM practice was assessed using spatiotemporal kernel density estimation in R with spatstat and sparr packages. 
Results: Our database comprised 743 records, spanning 40 countries across five regions. We found the highest incidence of ICM throughout the New World and identified clusters of diverse modification types and potential areas of missing data due to sampling bias in the archaeological record. 
Discussion: We present a global overview of potential areas and times of diffusion, and cultural convergence. There were probably multiple developments of CM, mirrored also in the different social contexts and modes in which CM has been recorded.

Premium content

Strategic Social Leaning in the Era of Hypercompetent Artificial Intelligence: A Large-Scale, Longitudinal Case Study using the Game of Go

Downloads

Next from Cultural Evolution Society 2022

The evolution of false theories

Similar lecture

Multiagent Model-based Credit Assignment for Continuous Control

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES