Social judgment is a social inference process whereby an agent singles out individuals to blame or credit for multi-agent activities. Such inferences are a key aspect of social intelligence that underlie social planning, social learning, natural language pragmatics and computational models of emotion. With the advance of multi-agent interactive systems and the need of designing socially aware systems and interfaces to interact with people, it is increasingly important to model this human-centric form of (...) class='Hi'>social inference. Based on psychological attribution theory, this paper presents a general computational framework to automate social inference based on an agent’s causal knowledge and observations of interaction. (shrink)

We incorporate social reasoning about groups of informants into a model of word learning, and show that the model accounts for infant looking behavior in tasks of both word learning and recognition. Simulation 1 models an experiment where 16-month-old infants saw familiar objects labeled either correctly or incorrectly, by either adults or audio talkers. Simulation 2 reinterprets puzzling data from the Switch task, an audiovisual habituation procedure wherein infants are tested on familiarized associations between novel objects and labels. Eight-month-olds (...) outperform 14-month-olds on the Switch task when required to distinguish labels that are minimal pairs, but 14-month-olds' performance is improved by habituation stimuli featuring multiple talkers. Our modeling results support the hypothesis that beliefs about knowledgeability and group membership guide infant looking behavior in both tasks. These results show that social and linguistic development interact in non-trivial ways, and that social categorization findings in developmental psychology could have substantial implications for understanding linguistic development in realistic settings where talkers vary according to observable features correlated with social groupings, including linguistic, ethnic, and gendered groups. (shrink)

When making inferences about similar others, people anchor and adjust away from themselves. However, research on relational self theory suggests the possibility of using knowledge about others as an anchor when they are more similar to a target. We investigated whether social inferences are made on the basis of significant other knowledge through an anchoring and adjustment process, and whether anchoring on a significant other is more effortful than anchoring on the self. Participants answered questions about their likes and (...) habits, as well as the likes and habits of a significant other, a target similar to their significant other, and a yoked control. We found that prediction differences between the significant other and similar target led to longer response times, and we found the opposite effect for self and target differences, suggesting anchoring and adjustment from the significant other rather than the self. These effects were moderated by the source-relative salience of the dimension being evaluated. The evidence was mixed with respect to the question of whether anchoring on a significant other is more effortful than anchoring on the self. (shrink)

We distinguish between three philosophical views on the neuroscience of predictive models: predictive coding, predictive processing and predictive engagement. We examine the concept of active inference under each model and then ask how this concept informs discussions of social cognition. In this context we consider Frith and Friston’s proposal for a neural hermeneutics, and we explore the alternative model of enactivist hermeneutics.

In this paper, we introduce an ecological account of communication according to which acts of communication are active inferences achieved by affecting the behavior of a target organism via the modification of its field of affordances. Constraining a target organism’s behavior constitutes a mechanism of socially extended active inference, allowing organisms to proactively regulate their inner states through the behavior of other organisms. In this general conception of communication, the type of cooperative communication characteristic of human communicative interaction is (...) a way of constraining interaction dynamics toward the goals of a given joint action by constructing and altering shared fields of affordances. This account embraces a pragmatist view according to which communication is a form of action aiming to influence the behavior of a target, and stands against the traditional transmission view according to which communication fundamentally serves to convey information. Understanding acts of communication as active inference under an ecological interpretation allows us to link communicative and ultimately linguistic behavior to the biological imperative of minimizing free energy and to emphasize the action-oriented nature of communicative interaction. (shrink)

How is metaphysics related to the empirical sciences? Should metaphysics in general be guided by the sources, methods and results of the sciences? And what about the special case of the metaphysics of the social world: should it likewise be guided by the sources, methods and results of the social sciences? In her paper “Social Science as a Guide to Social Metaphysics?”, K. Hawley raises the question: If we are sympathetic to the project of naturalising metaphysics, (...) how should we approach the metaphysics of the social world? She proceeds by discussing three approaches to social metaphysics: inference to the best explanation from current social science, descriptive conceptual analysis, and normative, especially ‘ameliorative’ projects. At the end of her discussion, she reaches a rather pessimistic conclusion, especially as regards the IBE approach: “a number of phenomena indicate that the prospects for securely basing social metaphysics via inference to the best explanation from social science are currently faint. […] We need to look elsewhere if we are to develop a metaphysics of the social world.” In my comments on her paper, I try to re-animate the program of an inductive metaphysics by defending the idea that the method of inference to the best explanation should be the central method of justification for metaphysics in general and for social metaphysics in particular. (shrink)

Several authors have claimed that mechanisms play a vital role in distinguishing between causation and mere correlation in the social sciences. Such claims are sometimes interpreted to mean that without mechanisms, causal inference in social science is impossible. The author agrees with critics of this proposition but explains how the account of how mechanisms aid causal inference can be interpreted in a way that does not depend on it. Nevertheless, he shows that this more charitable version (...) of the account is still unsuccessful as it stands. Consequently, he advances a proposal for shoring up the account, which is founded on the possibility of acquiring knowledge of social mechanisms by linking together norms or practices found in a society. Key Words: causality • social mechanisms • interpretation • anthropology. (shrink)

This article compares inference to the best explanation with Bayes’s rule in a social setting, specifically, in the context of a variant of the Hegselmann–Krause model in which agents not only update their belief states on the basis of evidence they receive directly from the world, but also take into account the belief states of their fellow agents. So far, the update rules mentioned have been studied only in an individualistic setting, and it is known that in such (...) a setting both have their strengths as well as their weaknesses. It is shown here that in a social setting, inference to the best explanation outperforms Bayes’s rule according to every desirable criterion. 1 What Is Inference to the Best Explanation?2 Judging the Rules—By Which Lights?3 From an Individualistic to a Social Perspective 3.1 The Hegselmann–Krause model 3.2 A probabilistic extension of the Hegselmann–Krause model 3.3 Simulations4 Results and Discussion5 Interpretation6 Conclusion. (shrink)

The paper has two aims. First, to show that we need social mechanisms to establish the policy relevance of causal claims, even if it is possible to build a good argument for those claims without knowledge of mechanisms. Second, to show that although social scientists can, in principle, do without social mechanisms when they argue for causal claims, in reality scientific practice contexts where they do not need mechanisms are very rare. Key Words: social mechanisms • (...) causal inference • social policy. (shrink)

This release includes the data and code used in: Malaterre, C., F. Lareau (2023) Inferring social networks from unstructured text data: A proof of concept detection of “hidden communities of interest”. Text and Data Analytics for Policy.

Many human cultural traits become increasingly beneficial as they are repeatedly transmitted, thanks to an accumulation of modifications made by successive generations. But how do later generations typically avoid modifications which revert traits to less beneficial forms already sampled and rejected by earlier generations? And how can later generations do so without direct exposure to their predecessors' behavior? One possibility is that learners are sensitive to cues of non‐random production in others' behavior, and that particular variants (e.g., those containing structural (...) regularities unlikely to occur spontaneously) have been produced deliberately and with some effort. If this non‐random behavior is attributed to an informed strategy, then the learner may infer that apparent avoidance of certain possibilities indicates that these have already been sampled and rejected. This could potentially prevent performance plateaus resulting from learners modifying inherited behaviors randomly. We test this hypothesis in four experiments in which participants, either individually or in interacting dyads, attempt to locate rewards in a search grid, guided by partial information about another individual's experience of the task. We find that in some contexts, valid inferences about another's behavior can be made from partial information, and these inferences can be used in a way which facilitates trait adaptation. However, the benefit of these inferences appears to be limited, and in many contexts—including some which have the potential to make inferring the experience of another individual easier—there appears to be no benefit at all. We suggest that inferring previous behavior from partial social information plays a minimal role in the adaptation of cultural traits. (shrink)

This commentary is in agreement with the thrust of Koehler's target article. The issue I deal with is whether a Bayesian framework represents an adequate general normative framework for deciding the rationality of lay judgments, even when it can be unambiguously applied.

The article discusses the problem of the naturalistic methodology of social ontology. Following Katherine Hawley's (2018) analysis, the author considers three approaches: conceptual analysis, the ameliorative (or normative) approach, and inference to the best explanation (from best social science to social ontology). Hawley concludes that only the first two can provide a viable naturalistic social metaphysics, and the latter cannot. The author, drawing on the notion of naturalistic limitations of social ontology, shows that only (...) a conclusion to the best explanation can lead to a consistent realistic social ontology; this methodology avoids the problem of being rooted in concepts that do not have an empirical basis, which contradicts realism, and also does not imply normative prescription, which also contradicts the realistic description of the world. The problem critics attribute to inference to the best explanation - the lack of a predictively strong theoretical core of social theory from which to infer social ontology - can be resolved by satisfying three criteria (Turner, 2007): physical realizability, computability, and cognitive realism. The author proposes to implement them through the introduction of naturalistic restrictions into game theory. Formal models of coordination described by a correlated equilibrium - coupled with evolutionary explanations of the cognitive mechanisms responsible for the causal processes that make formally described coordination possible - may provide the desired theoretical core from which ontology can be derived. (shrink)

This chapter explores the idea that the need to establish common knowledge is one feature that makes social cognition stand apart in important ways from cognition in general. We develop this idea on the background of the claim that social cognition is nothing but a type of causal inference. We focus on autism as our test-case, and propose that a specific type of problem with common knowledge processing is implicated in challenges to social cognition in autism (...) spectrum disorder (ASD). This problem has to do with the individual’s assessment of the reliability of messages that are passed between people as common knowledge emerges. The proposal is developed on the background of our own empirical studies and outlines different ways common knowledge might be comprised. We discuss what these issues may tell us about ASD, about the relation between social and non-social cognition, about social objects, and about the dynamics of social networks. (shrink)

In this paper I defend interaction theory (IT) as an alternative to both theory theory (TT) and simulation theory (ST). IT opposes the basic suppositions that both TT and ST depend upon. I argue that the various capacities for primary and secondary intersubjectivity found in infancy and early childhood should not be thought of as precursors to later developing capacities for using folk psychology or simulation routines. They are not replaced or displaced by such capacities in adulthood, but rather continue (...) to operate as our ordinary and everyday basis for social cognition. I also argue that enactive perception rather than implicit simulation is the best model for explaining these capacities. (shrink)

This aim of this paper is two-fold: it critically analyses and rejects accounts blending active inference as theory of mind and enactivism; and it advances an enactivist-dynamic understanding of social cognition that is compatible with active inference. While some social cognition theories seemingly take an enactive perspective on social cognition, they explain it as the attribution of mental states to other people, by assuming representational structures, in line with the classic Theory of Mind. Holding both (...) enactivism and ToM, we argue, entails contradiction and confusion due to two ToM assumptions widely known to be rejected by enactivism: that social cognition reduces to mental representation and social cognition is a hardwired contentful ‘toolkit’ or ‘starter pack’ that fuels the model-like theorising supposed in. The paper offers a positive alternative, one that avoids contradictions or confusion. After rejecting ToM-inspired theories of social cognition and clarifying the profile of social cognition under enactivism, that is without assumptions and, the last section advances an enactivist-dynamic model of cognition as dynamic, real-time, fluid, contextual social action, where we use the formalisms of dynamical systems theory to explain the origins of socio-cognitive novelty in developmental change and active inference as a tool to demonstrate social understanding as generalised synchronisation. (shrink)

Language provides rich social information about its speakers. For instance, adults and children make inferences about a speaker's social identity, geographic origins, and group membership based on her language and accent. Although infants prefer speakers of familiar languages, little is known about the developmental origins of humans’ sensitivity to language as marker of social identity. We investigated whether 9-month-olds use the language a person speaks as an indicator of that person's likely social relationships. Infants were familiarized (...) with videos of two people who spoke the same or different languages, and then viewed test videos of those two individuals affiliating or disengaging. Results suggest that infants expected two people who spoke the same language to be more likely to affiliate than two people who spoke different languages. Thus, infants view language as a meaningful social marker and use language to make inferences about third-party social relationships. (shrink)

Generics (e.g., “Ravens are black”) express generalizations about categories or their members. Previous research found that generics about animals are interpreted as broadly true of members of a kind, yet also accepted based on minimal evidence. This asymmetry is important for suggesting a mechanism by which unfounded generalizations may flourish; yet, little is known whether this finding extends to generics about groups of people (heretofore, “social generics”). Accordingly, in four preregistered studies (n = 665), we tested for an inferential (...) asymmetry for generics regarding novel groups of animals versus people. Participants were randomly assigned to either an Implied Prevalence task (given a generic, asked to estimate the prevalence of a property) or a Truth-Conditions task (given prevalence information, asked whether a generic was true or false). A generic asymmetry was found in both domains, at equivalent levels. The asymmetry also extended to properties varying in valence (dangerous and neutral). Finally, there were differences as a function of property valence in the Implied Prevalence task and a small but consistent interaction between domain and prevalence in the Truth-Conditions task. We discuss the implications of these results for the semantics of generics, theoretical accounts of the asymmetry, and the relation between generics and stereotyping. (shrink)

The book is an epistemological monograph written from a multidisciplinary perspective. It provides a complex and realistic picture of cognition and rationality, as endowments aimed at making sense and reacting smartly to one's environment, be it epistemic, social or simply ecological. The first part of the book analyzes scientific modeling as products of the biological necessity to cope with the environment and be able to draw as many inferences as possible about it. Moreover, it develops an epistemological framework which (...) will be exploited in both the second part of the book, focusing on social cognition and cognitive niche construction, and the third part, dealing with the apparent irrationality of magical and religious belief. The book also discusses how both social networking and online religion influence cognition, rationality and irrational belief. (shrink)

All univocal analyses of causation face counterexamples. An attractive response to this situation is to become a pluralist about causal relationships. "Causal pluralism" is itself, however, a pluralistic notion. In this article, I argue in favor of pluralism about concepts of cause in the social sciences. The article will show that evidence for, inference from, and the purpose of causal claims are very closely linked. Key Words: causation • pluralism • evidence • methodology.

The notion that emotional expressions regulate social life by providing information is gaining popularity. Prior research on the effects of emotional expressions on observers’ inferential processes has focused mostly on inferences regarding the personality traits of the expresser, such as dominance and affiliation. We extend this line of research by exploring the possibility that emotional expressions shape observers’ construal of social situations. Across three vignette studies, an interaction partner's expressions of anger, compared to expressions of happiness or disappointment, (...) led observers to construe hypothetical situations as less cooperative, both in dyads and groups. These effects occurred even when factual information regarding the cooperativeness or competitiveness of the situation was provided, attesting to the power of emotional expressions in shaping the construal of social situations. Results are discussed in relation to appraisal theory, reverse appraisals, emotions as social information theory, and the emergence of cooperation in groups and cultures. (shrink)

Grünbaum argues that psychoanalysis cannot justify its inferences regarding motives using its own methodology, as only the employment of Mill’s canons can justify causal inferences (which inferences to motives are). I consider an argument offered by Hopkins regarding the nature and status of our everyday inferences from other people’s behavior to their motives that seeks to rebut Grünbaum’s charge by defending a form of inference to the best explanation that makes use of connections in intentional content between behavior and (...) motives. I argue that Hopkins succeeds in defeating Grünbaum’s objection as it is presented, but that work in social psychology presents a further challenge. I discuss the extent to which the challenge can be met, and conclude that certain types of inference in psychoanalysis are justifiable, but others, including those which are the target of Grünbaum’s objection, cannot be justified by the methods defended by Hopkins. (shrink)

In two studies, we predicted and found that inferences about motive and character influence intentionality attributions about foreseeable consequences of action (i.e., side effects). First, we show that inferences about intentionality are greater for good side effects than bad side effects when a target person's character is described positively. In Study 2, we manipulated information about a target person and found that inferences about intentionality were greater when side effects were consistent with a target person's character and motives. Overall, our (...) data cast doubt on the generality of the side-effect effect. We discuss our findings and their implications for future research on intentionality and social perception. (shrink)

In this book, Reiss argues in favor of a tight fit between evidence, concept and purpose in our causal investigations in the sciences. There is no doubt that the sciences employ a vast array of techniques to address causal questions such as controlled experiments, randomized trials, statistical and econometric tools, causal modeling and thought experiments. But how do these different methods relate to each other and to the causal inquiry at hand? Reiss argues that there is no "gold standard" in (...) settling causal issues against which other methods can be measured. Rather, the various methods of inference tend to be good only relative to certain interpretations of the word "cause", and each interpretation, in turn, helps to address some salient purpose but not others. The main objective of this book is to explore the metaphysical and methodological consequences of this view in the context of numerous cases studies from the natural and social sciences. (shrink)

The problem of inferring probability comparisons between events from an initial set of comparisons arises in several contexts, ranging from decision theory to artificial intelligence to formal semantics. In this paper, we treat the problem as follows: beginning with a binary relation ≥ on events that does not preclude a probabilistic interpretation, in the sense that ≥ has extensions that are probabilistically representable, we characterize the extension ≥+ of ≥ that is exactly the intersection of all probabilistically representable extensions of (...) ≥. This extension ≥+ gives us all the additional comparisons that we are entitled to infer from ≥, based on the assumption that there is some probability measure of which ≥ gives us partial qualitative information. We pay special attention to the problem of extending an order on states to an order on events. In addition to the probabilistic interpretation, this problem has a more general interpretation involving measurement of any additive quantity: e.g., given comparisons between the weights of individual objects, what comparisons between the weights of groups of objects can we infer? (shrink)

Social and linguistic perceptions are linked. On one hand, talker identity affects speech perception. On the other hand, speech itself provides information about a talker's identity. Here, we propose that the same probabilistic knowledge might underlie both socially conditioned linguistic inferences and linguistically conditioned social inferences. Our computational–level approach—the ideal adapter—starts from the idea that listeners use probabilistic knowledge of covariation between social, linguistic, and acoustic cues in order to infer the most likely explanation of the speech (...) signals they hear. As a first step toward understanding social inferences in this framework, we use a simple ideal observer model to show that it would be possible to infer aspects of a talker's identity using cue distributions based on actual speech production data. This suggests the possibility of a single formal framework for social and linguistic inferences and the interactions between them. (shrink)

One of the main challenges for computational legal research is drawing up innovative heuristics to derive actionable knowledge from legal documents. While a large part of the research has been so far devoted to the extraction of purely legal information, less attention has been paid to seeking out in the texts the clues of more complex entities: legally relevant facts whose detection requires to link and interpret, as a unified whole, legal information and results of empirical analyses. This paper presents (...) an ongoing research that points in this direction, trying to devise new ways to support public prosecutors in assessing the dangerousness of individuals and groups under investigation, an activity that precisely relies on the cross-sectional evaluation of legal and empirical data. A knowledge mining strategy will be outlined that lines up, into a single metaheuristic model, information extraction, network-based inference, machine learning and visual analytics. We will focus, in particular, on the integration of graph-based inference and machine learning methods used both to support classification tasks and to explore new forms of man-machine cooperation. Experiments made involving public prosecutors from the Italian Anti-Mafia Investigation Directorate and using data from real investigations have not only shown the potentialities of our approach but also offered an opportunity to reflect on the role we could assign to AI when thinking about the future of legal science and practice. (shrink)

Transitive inference (TI) refers to social cognition that facilitates the discernment of unknown relationships between individuals using known relationships. It is extensively reported that TI evolves in animals living in a large group because TI could assess relative rank without deducing all dyadic relationships, which averts costly fights. The relationships in a large group become so complex that social cognition may not be developed adequately to handle such complexity. If members apply TI to all possible members in (...) the group, TI requires extremely highly developed cognitive abilities especially in a large group. Instead of developing cognitive abilities significantly, animals may apply simplified TI we call reference TI in this study as heuristic approaches. The reference TI allows members to recognize and remember social interactions only among a set of reference members rather than all potential members. Our study assumes that information processes in the reference TI comprises (1) the number of reference members based on which individuals infer transitively, (2) the number of reference members shared by the same strategists, and (3) memory capacity. We examined how information processes evolve in a large group using evolutionary simulations in the hawk–dove game. Information processes with almost any numbers of reference members could evolve in a large group as long as the numbers of shared reference member are high because information from the others’ experiences is shared. TI dominates immediate inference, which assesses relative rank on direct interactions, because TI could establish social hierarchy more rapidly applying information from others’ experiences. (shrink)

An accessible and contemporary introduction to the methods for determining cause and effect in the social sciences Causal inference encompasses the tools that allow social scientists to determine what causes what. Economists--who generally can't run controlled experiments to test and validate their hypotheses--apply these tools to observational data to make connections. In a messy world, causal inference is what helps establish the causes and effects of the actions being studied, whether the impact (or lack thereof) of (...) increases in the minimum wage on employment, the effects of early childhood education on incarceration later in life, or the introduction of malaria nets in developing regions on economic growth. Scott Cunningham introduces students and practitioners to the methods necessary to arrive at meaningful answers to the questions of causation, using a range of modeling techniques and coding instructions for both the R and Stata programming languages. - -. (shrink)

This article examines two theses formulated by Russo and Williamson in their study of causal inference in the health sciences. The two theses are assessed against evidence from a specific case in the social sciences, i.e., research on the institutional determinants of the aggregate unemployment rate. The first Russo–Williamson Thesis is that a causal claim can only be established when it is jointly supported by difference-making and mechanistic evidence. This thesis is shown not to hold. While researchers in (...) my case study draw extensively on both types of evidence, one causal claim out of the three analyzed is established even though it is exclusively supported by mechanistic evidence. The second Russo–Williamson Thesis is that standard accounts of causality fail to handle the dualist epistemology highlighted in the first Thesis. I argue that a counterfactual-manipulationist account of causality—which is endorsed by many philosophers as well as many social scientists—can perfectly make sense of the typical strategy in my case study to draw on both difference-making and mechanistic evidence; it is just an instance of the common strategy of increasing evidential variety. (shrink)

Mainstream social psychology focuses on how people characteristically violate norms of action through social misbehaviors such as conformity with false majority judgments, destructive obedience, and failures to help those in need. Likewise, they are seen to violate norms of reasoning through cognitive errors such as misuse of social information, self-enhancement, and an over-readiness to attribute dispositional characteristics. The causes of this negative research emphasis include the apparent informativeness of norm violation, the status of good behavior and judgment (...) as unconfirmable null hypotheses, and the allure of counter-intuitive findings. The shortcomings of this orientation include frequently erroneous imputations of error, findings of mutually contradictory errors, incoherent interpretations of error, an inability to explain the sources of behavioral or cognitive achievement, and the inhibition of generalized theory. Possible remedies include increased attention to the complete range of behavior and judgmental accomplishment, analytic reforms emphasizing effect sizes and Bayesian inference, and a theoretical paradigm able to account for both the sources of accomplishment and of error. A more balanced social psychology would yield not only a more positive view of human nature, but also an improved understanding of the bases of good behavior and accurate judgment, coherent explanations of occasional lapses, and theoretically grounded suggestions for improvement. Key Words: Bayesian inference; biases; normative models; personality; positive psychology; rationality; reasoning; social behavior; social judgment; social psychology. (shrink)

The social psychology of science is a compelling new area of study whose shape is still emerging. This erudite and innovative book outlines a theoretical and methodological agenda for this new field, and bridges the gap between the individually focused aspects of psychology and the sociological elements of science studies. Presenting a side of social psychology that, until now, has received almost no attention in the social sciences literature, this volume offers the first detailed and comprehensive study (...) of the social psychology of science, complete with a large number of empirical and theoretical examples. The volume's introductory section provides a detailed analysis of how modern social psychology might apply to the study of science. Chapters show how to analyze science in terms of social cognition, attribution theory, attitudes and attitude change, social motivation, social influence and social conformity, and intergroup relations, weaving extensive illustrations from the science studies literature into the theoretical analysis. The nature and role of experimentation are discussed, as are metaanalytic methods for summarizing the results of multiple studies. Ways to facilitate the generalization of causal inferences from experimental work are also examined. The book focuses on such topics as interactions among small groups of scientists, and the impact of social motivation, influence, and conformity on scientific work. Also covered are scientists' responses to ethical issues in research, differences in cognitive style distribution, creativity in research and development, and the sociologists's view of the social psychology of science and technology. In addition, the book provides two annotated bibliographies, one on the philosophy of science and the other on social psychology, to guide readers in both disciplines to salient recent works. Valuable to the entire science studies community, this text will be of special interest to philosophers, sociologists, psychologists, and historians of science interested in the nature of knowledge development in science. Because of its novel application of social psychological theories and methods, this book will be useful as a primary text or a secondary text in courses on science studies in psychology, sociology, or philosophy departments. (shrink)

Collaboration requires agents to coordinate their behavior on the fly, sometimes cooperating to solve a single task together and other times dividing it up into sub‐tasks to work on in parallel. Underlying the human ability to collaborate is theory‐of‐mind (ToM), the ability to infer the hidden mental states that drive others to act. Here, we develop Bayesian Delegation, a decentralized multi‐agent learning mechanism with these abilities. Bayesian Delegation enables agents to rapidly infer the hidden intentions of others by inverse planning. (...) We test Bayesian Delegation in a suite of multi‐agent Markov decision processes inspired by cooking problems. On these tasks, agents with Bayesian Delegation coordinate both their high‐level plans (e.g., what sub‐task they should work on) and their low‐level actions (e.g., avoiding getting in each other's way). When matched with partners that act using the same algorithm, Bayesian Delegation outperforms alternatives. Bayesian Delegation is also a capable ad hoc collaborator and successfully coordinates with other agent types even in the absence of prior experience. Finally, in a behavioral experiment, we show that Bayesian Delegation makes inferences similar to human observers about the intent of others. Together, these results argue for the centrality of ToM for successful decentralized multi‐agent collaboration. (shrink)

To learn about real world phenomena, scientists have traditionally used models with clearly interpretable elements. However, modern machine learning (ML) models, while powerful predictors, lack this direct elementwise interpretability (e.g. neural network weights). Interpretable machine learning (IML) offers a solution by analyzing models holistically to derive interpretations. Yet, current IML research is focused on auditing ML models rather than leveraging them for scientific inference. Our work bridges this gap, presenting a framework for designing IML methods—termed ’property descriptors’—that illuminate not (...) just the model, but also the phenomenon it represents. We demonstrate that property descriptors, grounded in statistical learning theory, can effectively reveal relevant properties of the joint probability distribution of the observational data. We identify existing IML methods suited for scientific inference and provide a guide for developing new descriptors with quantified epistemic uncertainty. Our framework empowers scientists to harness ML models for inference, and provides directions for future IML research to support scientific understanding. (shrink)

As social beings, people need to be able to interact intelligently with others in their social environment. Accordingly, people spend much time conversing with one another in order to understand the broad and fine aspects of the relations that link them. They are especially interested in the interactive behaviors that constitute social relations, such as mutual aid, gift giving and exchange, sharing, informal socializing, or deception. The evaluations of these behaviors are embedded in social relationships and (...) charged with values and emotions. We developed tasks to probe how people in an unfamiliar socio-cultural setting understand and account for the behavior of others conditional upon their category membership – by trying to elicit the basic categories, stereotypes, and models that inform the causal perceptions, inferences and reasoning people use in understanding others’ interactive behaviors – and we tested these tasks among the Wampar in Papua New Guinea. The results show changes in the relevance of social categories among the Wampar but also, and perhaps more important, limitations in the translation and applicability of cognitive tasks. (shrink)