People construct simplified mental representations to plan.

One of the most striking features of human cognition is the ability to plan. Two aspects of human planning stand out-its efficiency and flexibility. Efficiency is especially impressive because plans must often be made in complex environments, and yet people successfully plan solutions to many everyday problems despite having limited cognitive resources1-3. Standard accounts in psychology, economics and artificial intelligence have suggested that human planning succeeds because people have a complete representation of a task and then use heuristics to plan future actions in that representation4-11. However, this approach generally assumes that task representations are fixed. Here we propose that task representations can be controlled and that such control provides opportunities to quickly simplify problems and more easily reason about them. We propose a computational account of this simplification process and, in a series of preregistered behavioural experiments, show that it is subject to online cognitive control12-14 and that people optimally balance the complexity of a task representation and its utility for planning and acting. These results demonstrate how strategically perceiving and conceiving problems facilitates the effective use of limited cognitive resources.

Integrating explanation and prediction in computational social science.

Computational social science is more than just large repositories of digital data and the computational methods needed to construct and analyse them. It also represents a convergence of different fields with different ways of thinking about and doing science. The goal of this Perspective is to provide some clarity around how these approaches differ from one another and to propose how they might be productively integrated. Towards this end we make two contributions. The first is a schema for thinking about research activities along two dimensions-the extent to which work is explanatory, focusing on identifying and estimating causal effects, and the degree of consideration given to testing predictions of outcomes-and how these two priorities can complement, rather than compete with, one another. Our second contribution is to advocate that computational social scientists devote more attention to combining prediction and explanation, which we call integrative modelling, and to outline some practical suggestions for realizing this goal.

Superhuman artificial intelligence can improve human decision-making by increasing novelty.

How will superhuman artificial intelligence (AI) affect human decision-making? And what will be the mechanisms behind this effect? We address these questions in a domain where AI already exceeds human performance, analyzing more than 5.8 million move decisions made by professional Go players over the past 71 y (1950 to 2021). To address the first question, we use a superhuman AI program to estimate the quality of human decisions across time, generating 58 billion counterfactual game patterns and comparing the win rates of actual human decisions with those of counterfactual AI decisions. We find that humans began to make significantly better decisions following the advent of superhuman AI. We then examine human players' strategies across time and find that novel decisions (i.e., previously unobserved moves) occurred more frequently and became associated with higher decision quality after the advent of superhuman AI. Our findings suggest that the development of superhuman AI programs may have prompted human players to break away from traditional strategies and induced them to explore novel moves, which in turn may have improved their decision-making.

Toward a Rational and Mechanistic Account of Mental Effort.

In spite of its familiar phenomenology, the mechanistic basis for mental effort remains poorly understood. Although most researchers agree that mental effort is aversive and stems from limitations in our capacity to exercise cognitive control, it is unclear what gives rise to those limitations and why they result in an experience of control as costly. The presence of these control costs also raises further questions regarding how best to allocate mental effort to minimize those costs and maximize the attendant benefits. This review explores recent advances in computational modeling and empirical research aimed at addressing these questions at the level of psychological process and neural mechanism, examining both the limitations to mental effort exertion and how we manage those limited cognitive resources. We conclude by identifying remaining challenges for theoretical accounts of mental effort as well as possible applications of the available findings to understanding the causes of and potential solutions for apparent failures to exert the mental effort required of us.

Deep models of superficial face judgments.

The diversity of human faces and the contexts in which they appear gives rise to an expansive stimulus space over which people infer psychological traits (e.g., trustworthiness or alertness) and other attributes (e.g., age or adiposity). Machine learning methods, in particular deep neural networks, provide expressive feature representations of face stimuli, but the correspondence between these representations and various human attribute inferences is difficult to determine because the former are high-dimensional vectors produced via black-box optimization algorithms. Here we combine deep generative image models with over 1 million judgments to model inferences of more than 30 attributes over a comprehensive latent face space. The predictive accuracy of our model approaches human interrater reliability, which simulations suggest would not have been possible with fewer faces, fewer judgments, or lower-dimensional feature representations. Our model can be used to predict and manipulate inferences with respect to arbitrary face photographs or to generate synthetic photorealistic face stimuli that evoke impressions tuned along the modeled attributes.

Leveraging artificial intelligence to improve people's planning strategies.

SignificanceMany bad decisions and their devastating consequences could be avoided if people used optimal decision strategies. Here, we introduce a principled computational approach to improving human decision making. The basic idea is to give people feedback on how they reach their decisions. We develop a method that leverages artificial intelligence to generate this feedback in such a way that people quickly discover the best possible decision strategies. Our empirical findings suggest that a principled computational approach leads to improvements in decision-making competence that transfer to more difficult decisions in more complex environments. In the long run, this line of work might lead to apps that teach people clever strategies for decision making, reasoning, goal setting, planning, and goal achievement.

Improving Memory Search Through Model-Based Cue Selection.

We often use cues from our environment when we get stuck searching our memories, but prior research has failed to show benefits of cuing with other, randomly selected list items during memory search. What accounts for this discrepancy? We proposed that cues' content critically determines their effectiveness and sought to select the right cues by building a computational model of how cues affect memory search. Participants (N = 195 young adults from the United States) recalled significantly more items when receiving our model's best (vs. worst) cue. Our model provides an account of why some cues better aid recall: Effective cues activate contexts most similar to the remaining items' contexts, facilitating recall in an unsearched area of memory. We discuss our contributions in relation to prominent theories about the effect of external cues.

Disentangling Abstraction from Statistical Pattern Matching in Human and Machine Learning.

The ability to acquire abstract knowledge is a hallmark of human intelligence and is believed by many to be one of the core differences between humans and neural network models. Agents can be endowed with an inductive bias towards abstraction through meta-learning, where they are trained on a distribution of tasks that share some abstract structure that can be learned and applied. However, because neural networks are hard to interpret, it can be difficult to tell whether agents have learned the underlying abstraction, or alternatively statistical patterns that are characteristic of that abstraction. In this work, we compare the performance of humans and agents in a meta-reinforcement learning paradigm in which tasks are generated from abstract rules. We define a novel methodology for building "task metamers" that closely match the statistics of the abstract tasks but use a different underlying generative process, and evaluate performance on both abstract and metamer tasks. We find that humans perform better at abstract tasks than metamer tasks whereas common neural network architectures typically perform worse on the abstract tasks than the matched metamers. This work provides a foundation for characterizing differences between humans and machine learning that can be used in future work towards developing machines with more human-like behavior.

Humans decompose tasks by trading off utility and computational cost.

Human behavior emerges from planning over elaborate decompositions of tasks into goals, subgoals, and low-level actions. How are these decompositions created and used? Here, we propose and evaluate a normative framework for task decomposition based on the simple idea that people decompose tasks to reduce the overall cost of planning while maintaining task performance. Analyzing 11,117 distinct graph-structured planning tasks, we find that our framework justifies several existing heuristics for task decomposition and makes predictions that can be distinguished from two alternative normative accounts. We report a behavioral study of task decomposition (N = 806) that uses 30 randomly sampled graphs, a larger and more diverse set than that of any previous behavioral study on this topic. We find that human responses are more consistent with our framework for task decomposition than alternative normative accounts and are most consistent with a heuristic-betweenness centrality-that is justified by our approach. Taken together, our results suggest the computational cost of planning is a key principle guiding the intelligent structuring of goal-directed behavior.

Serial reproduction reveals the geometry of visuospatial representations.

An essential function of the human visual system is to locate objects in space and navigate the environment. Due to limited resources, the visual system achieves this by combining imperfect sensory information with a belief state about locations in a scene, resulting in systematic distortions and biases. These biases can be captured by a Bayesian model in which internal beliefs are expressed in a prior probability distribution over locations in a scene. We introduce a paradigm that enables us to measure these priors by iterating a simple memory task where the response of one participant becomes the stimulus for the next. This approach reveals an unprecedented richness and level of detail in these priors, suggesting a different way to think about biases in spatial memory. A prior distribution on locations in a visual scene can reflect the selective allocation of coding resources to different visual regions during encoding ("efficient encoding"). This selective allocation predicts that locations in the scene will be encoded with variable precision, in contrast to previous work that has assumed fixed encoding precision regardless of location. We demonstrate that perceptual biases covary with variations in discrimination accuracy, a finding that is aligned with simulations of our efficient encoding model but not the traditional fixed encoding view. This work demonstrates the promise of using nonparametric data-driven approaches that combine crowdsourcing with the careful curation of information transmission within social networks to reveal the hidden structure of shared visual representations.

Replies to commentaries on beyond playing 20 questions with nature.

Commentaries on the target article offer diverse perspectives on integrative experiment design. Our responses engage three themes: (1) Disputes of our characterization of the problem, (2) skepticism toward our proposed solution, and (3) endorsement of the solution, with accompanying discussions of its implementation in existing work and its potential for other domains. Collectively, the commentaries enhance our confidence in the promise and viability of integrative experiment design, while highlighting important considerations about how it is used.

Rational Simplification and Rigidity in Human Planning.

Planning underpins the impressive flexibility of goal-directed behavior. However, even when planning, people can display surprising rigidity in how they think about problems (e.g., "functional fixedness") that lead them astray. How can our capacity for behavioral flexibility be reconciled with our susceptibility to conceptual inflexibility? We propose that these tendencies reflect avoidance of two cognitive costs: the cost of representing task details and the cost of switching between representations. To test this hypothesis, we developed a novel paradigm that affords participants opportunities to choose different families of simplified representations to plan. In two preregistered, online studies (Ns = 377 and 294 adults), we found that participants' optimal behavior, suboptimal behavior, and reaction time were explained by a computational model that formalized people's avoidance of representational complexity and switching. These results demonstrate how the selection of simplified, rigid representations leads to the otherwise puzzling combination of flexibility and inflexibility observed in problem solving.

The pursuit of happiness: A reinforcement learning perspective on habituation and comparisons.

In evaluating our choices, we often suffer from two tragic relativities. First, when our lives change for the better, we rapidly habituate to the higher standard of living. Second, we cannot escape comparing ourselves to various relative standards. Habituation and comparisons can be very disruptive to decision-making and happiness, and till date, it remains a puzzle why they have come to be a part of cognition in the first place. Here, we present computational evidence that suggests that these features might play an important role in promoting adaptive behavior. Using the framework of reinforcement learning, we explore the benefit of employing a reward function that, in addition to the reward provided by the underlying task, also depends on prior expectations and relative comparisons. We find that while agents equipped with this reward function are less happy, they learn faster and significantly outperform standard reward-based agents in a wide range of environments. Specifically, we find that relative comparisons speed up learning by providing an exploration incentive to the agents, and prior expectations serve as a useful aid to comparisons, especially in sparsely-rewarded and non-stationary environments. Our simulations also reveal potential drawbacks of this reward function and show that agents perform sub-optimally when comparisons are left unchecked and when there are too many similar options. Together, our results help explain why we are prone to becoming trapped in a cycle of never-ending wants and desires, and may shed light on psychopathologies such as depression, materialism, and overconsumption.

Scaling up psychology via Scientific Regret Minimization.

Do large datasets provide value to psychologists? Without a systematic methodology for working with such datasets, there is a valid concern that analyses will produce noise artifacts rather than true effects. In this paper, we offer a way to enable researchers to systematically build models and identify novel phenomena in large datasets. One traditional approach is to analyze the residuals of models-the biggest errors they make in predicting the data-to discover what might be missing from those models. However, once a dataset is sufficiently large, machine learning algorithms approximate the true underlying function better than the data, suggesting, instead, that the predictions of these data-driven models should be used to guide model building. We call this approach "Scientific Regret Minimization" (SRM), as it focuses on minimizing errors for cases that we know should have been predictable. We apply this exploratory method on a subset of the Moral Machine dataset, a public collection of roughly 40 million moral decisions. Using SRM, we find that incorporating a set of deontological principles that capture dimensions along which groups of agents can vary (e.g., sex and age) improves a computational model of human moral judgment. Furthermore, we are able to identify and independently validate three interesting moral phenomena: criminal dehumanization, age of responsibility, and asymmetric notions of responsibility.

On the hazards of relating representations and inductive biases.

The success of models of human behavior based on Bayesian inference over logical formulas or programs is taken as evidence that people employ a "language-of-thought" that has similarly discrete and compositional structure. We argue that this conclusion problematically crosses levels of analysis, identifying representations at the algorithmic level based on inductive biases at the computational level.

A computational process-tracing method for measuring people's planning strategies and how they change over time.

One of the most unique and impressive feats of the human mind is its ability to discover and continuously refine its own cognitive strategies. Elucidating the underlying learning and adaptation mechanisms is very difficult because changes in cognitive strategies are not directly observable. One important domain in which strategies and mechanisms are studied is planning. To enable researchers to uncover how people learn how to plan, we offer a tutorial introduction to a recently developed process-tracing paradigm along with a new computational method for measuring the nature and development of a person's planning strategies from the resulting process-tracing data. Our method allows researchers to reveal experience-driven changes in people's choice of individual planning operations, planning strategies, strategy types, and the relative contributions of different decision systems. We validate our method on simulated and empirical data. On simulated data, its inferences about the strategies and the relative influence of different decision systems are accurate. When evaluated on human data generated using our process-tracing paradigm, our computational method correctly detects the plasticity-enhancing effect of feedback and the effect of the structure of the environment on people's planning strategies. Together, these methods can be used to investigate the mechanisms of cognitive plasticity and to elucidate how people acquire complex cognitive skills such as planning and problem-solving. Importantly, our methods can also be used to measure individual differences in cognitive plasticity and examine how different types (pedagogical) interventions affect the acquisition of cognitive skills.

The experimental evolution of human culture: flexibility, fidelity and environmental instability.

The past 2 Myr have seen both unprecedented environmental instability and the evolution of the human capacity for complex culture. This, along with the observation that cultural evolution occurs faster than genetic evolution, has led to the suggestion that culture is an adaptation to an unstable environment. We test this hypothesis by examining the ability of human social learning to respond to environmental changes. We do this by inserting human participants (n = 4800) into evolutionary simulations with a changing environment while varying the social information available to individuals across five conditions. We find that human social learning shows some signs of adaptation to environmental instability, including critical social learning, the adoption of up-and-coming traits and, unexpectedly, contrariness. However, these are insufficient to avoid significant fitness declines when the environment changes, and many individuals are highly conformist, which exacerbates the fitness effects of environmental change. We conclude that human social learning reflects a compromise between the competing needs for flexibility to accommodate environmental change and fidelity to accurately transmit valuable cultural information.

Globally Inaccurate Stereotypes Can Result From Locally Adaptive Exploration.

Inaccurate stereotypes-perceived differences among groups that do not actually differ-are prevalent and consequential. Past research explains stereotypes as emerging from a range of factors, including motivational biases, cognitive limitations, and information deficits. Considering the minimal forces required to produce inaccurate assumptions about group differences, we found that locally adaptive exploration is sufficient: An initial arbitrary interaction, if rewarding enough, may discourage people from investigating alternatives that would be equal or better. Historical accidents can snowball into globally inaccurate generalizations, and inaccurate stereotypes can emerge in the absence of real group differences. Using multiarmed-bandit models, we found that the mere act of choosing among groups with the goal of maximizing the long-term benefit of interactions is enough to produce inaccurate assessments of different groups. This phenomenon was reproduced in two large online experiments with English-speaking adults (N = 2,404), which demonstrated a minimal process that suffices to produce biased impressions.

Fixation patterns in simple choice reflect optimal information sampling.

Simple choices (e.g., eating an apple vs. an orange) are made by integrating noisy evidence that is sampled over time and influenced by visual attention; as a result, fluctuations in visual attention can affect choices. But what determines what is fixated and when? To address this question, we model the decision process for simple choice as an information sampling problem, and approximate the optimal sampling policy. We find that it is optimal to sample from options whose value estimates are both high and uncertain. Furthermore, the optimal policy provides a reasonable account of fixations and choices in binary and trinary simple choice, as well as the differences between the two cases. Overall, the results show that the fixation process during simple choice is influenced dynamically by the value estimates computed during the decision process, in a manner consistent with optimal information sampling.

Natural speech reveals the semantic maps that tile human cerebral cortex.

The meaning of language is represented in regions of the cerebral cortex collectively known as the 'semantic system'. However, little of the semantic system has been mapped comprehensively, and the semantic selectivity of most regions is unknown. Here we systematically map semantic selectivity across the cortex using voxel-wise modelling of functional MRI (fMRI) data collected while subjects listened to hours of narrative stories. We show that the semantic system is organized into intricate patterns that seem to be consistent across individuals. We then use a novel generative model to create a detailed semantic atlas. Our results suggest that most areas within the semantic system represent information about specific semantic domains, or groups of related concepts, and our atlas shows which domains are represented in each area. This study demonstrates that data-driven methods--commonplace in studies of human neuroanatomy and functional connectivity--provide a powerful and efficient means for mapping functional representations in the brain.