A rational account of pedagogical reasoning: Teaching by, and learning from, examples

Lexical Acquisition Through Category Matching: 12-Month-Old Infants Associate Words to Visual Categories

Although it is widely recognized that human infants build a sizeable conceptual repertoire before mastering language, it remains a matter of debate whether and to what extent early conceptual and category knowledge contributes to language development. We addressed this question by investigating whether 12-month-olds used preverbal categories to discover the meanings of new words. We showed that one group of infants ( n = 18) readily extended novel labels to previously unseen exemplars of preverbal visual categories after only a single labeling episode, but two other groups struggled to do so when taught labels for unfamiliar categories (those who had been previously exposed, n = 18, or not exposed, n = 18, to category tokens). These results suggest that infants expect labels to denote categories of objects and are equipped with learning mechanisms responsible for matching prelinguistic knowledge structures with linguistic inputs. This ability is consistent with the idea that our conceptual machinery provides building blocks for vocabulary and language acquisition.

The Emergence of Social Norms and Conventions

The utility of our actions frequently depends upon the beliefs and behavior of other agents. Thankfully, through experience, we learn norms and conventions that provide stable expectations for navigating our social world. Here, we review several distinct influences on their content and distribution. At the level of individuals locally interacting in dyads, success depends on rapidly adapting pre-existing norms to the local context. Hence, norms are shaped by complex cognitive processes involved in learning and social reasoning. At the population level, norms are influenced by intergenerational transmission and the structure of the social network. As human social connectivity continues to increase, understanding and predicting how these levels and time scales interact to produce new norms will be crucial for improving communities.

Means-Inference as a Source of Variability in Early Helping

Humans, as compared to their primate relatives, readily act on behalf of others: we help, inform, share resources with, and provide emotional comfort for others. Although these prosocial behaviors emerge early in life, some types of prosocial behaviors seem to emerge earlier than others, and some tasks elicit more reliable helping than others. Here we discuss existing perspectives on the sources of variability in early prosocial behaviors with a particular focus on the variability within the domain of instrumental helping. We suggest that successful helping behavior not only requires an understanding of others' goals (goal-inference), but also the ability to figure out how to help (means-inference). We review recent work that highlights two key factors that support means-inference: causal reasoning and sensitivity to the expected costs and rewards of actions. Once we begin to look closely at the process of deciding how to help someone, even a seemingly simple helping behavior is, in fact, a consequence of a sophisticated decision-making process; it involves reasoning about others (e.g., goals, actions, and beliefs), about the causal structure of the physical world, and about one's own ability to provide effective help. A finer-grained understanding of the role of these inferences may help explain the developmental trajectory of prosocial behaviors in early childhood. We discuss the promise of computational models that formalize this decision process and how this approach can provide additional insights into why humans show unparalleled propensity and flexibility in their ability to help others.

Unspoken Rules: Resolving Underdetermination With Closure Principles

When people learn normative systems, they do so based on limited evidence. Many of the possible actions that are available to an agent have never been explicitly permitted or prohibited. But people will often need to figure out whether those unspecified actions are permitted or prohibited. How does a learner resolve this incompleteness? The learner might assume if an action-type is not expressly forbidden, then acts of that type are permitted. This closure principle is one of Liberty. Alternatively, the learner might assume that if an action-type is not expressly permitted, then acts of that type are prohibited. This closure principle would be one of Residual Prohibition (Mikhail, 2011). On the basis of principles of pedagogical sampling (e.g., Shafto, Goodman, & Griffiths, ), we predicted that participants would infer the Liberty Principle (LP) when trained on prohibitions, and they would infer the Residual Prohibition Principle when trained on permissions. This is exactly what we found across several experiments. We also found a bias in favor of Liberty insofar as participants trained on both a prohibition and a permission rule tended still to infer the LP. However, we also found that if an action is potentially harmful, this diminishes the tendency to infer the LP.

The Theoretical and Methodological Opportunities Afforded by Guided Play With Young Children

For infants and young children, learning takes place all the time and everywhere. How children learn best both in and out of school has been a long-standing topic of debate in education, cognitive development, and cognitive science. Recently, guided play has been proposed as an integrative approach for thinking about learning as a child-led, adult-assisted playful activity. The interactive and dynamic nature of guided play presents theoretical and methodological challenges and opportunities. Drawing upon research from multiple disciplines, we discuss the integration of cutting-edge computational modeling and data science tools to address some of these challenges, and highlight avenues toward an empirically grounded, computationally precise and ecologically valid framework of guided play in early education.

Inductive reasoning 2.0

Inductive reasoning entails using existing knowledge to make predictions about novel cases. The first part of this review summarizes key inductive phenomena and critically evaluates theories of induction. We highlight recent theoretical advances, with a special emphasis on the structured statistical approach, the importance of sampling assumptions in Bayesian models, and connectionist modeling. A number of new research directions in this field are identified including comparisons of inductive and deductive reasoning, the identification of common core processes in induction and memory tasks and induction involving category uncertainty. The implications of induction research for areas as diverse as complex decision-making and fear generalization are discussed. This article is categorized under: Psychology > Reasoning and Decision Making Psychology > Learning.

Negotiating the Traffic: Can Cognitive Science Help Make Autonomous Vehicles a Reality?

To drive safely among human drivers, cyclists and pedestrians, autonomous vehicles will need to mimic, or ideally improve upon, humanlike driving. Yet, driving presents us with difficult problems of joint action: 'negotiating' with other users over shared road space. We argue that autonomous driving provides a test case for computational theories of social interaction, with fundamental implications for the development of autonomous vehicles.

A Diverse and Flexible Teaching Toolkit Facilitates the Human Capacity for Cumulative Culture

Human culture is uniquely complex compared to other species. This complexity stems from the accumulation of culture over time through high- and low-fidelity transmission and innovation. One possible reason for why humans retain and create culture, is our ability to modulate teaching strategies in order to foster learning and innovation. We argue that teaching is more diverse, flexible, and complex in humans than in other species. This particular characteristic of human teaching rather than teaching itself is one of the reasons for human's incredible capacity for cumulative culture. That is, humans unlike other species can signal to learners whether the information they are teaching can or cannot be modified. As a result teaching in humans can be used to support high or low fidelity transmission, innovation, and ultimately, cumulative culture.

Parameterizing developmental changes in epistemic trust

Children rely on others for much of what they learn, and therefore must track who to trust for information. Researchers have debated whether to interpret children's behavior as inferences about informants' knowledgeability only or as inferences about both knowledgeability and intent. We introduce a novel framework for integrating results across heterogeneous ages and methods. The framework allows application of a recent computational model to a set of results that span ages 8 months to adulthood and a variety of methods. The results show strong fits to specific findings in the literature trust, and correctly fails to fit one representative result from an adjacent literature. In the aggregate, the results show a clear development in children's reasoning about informants' intent and no appreciable changes in reasoning about informants' knowledgeability, confirming previous results. The results extend previous findings by modeling development over a much wider age range and identifying and explaining differences across methods.

Infant learning is influenced by local spurious generalizations

In previous work, 11-month-old infants were able to learn rules about the relation of the consonants in CVCV words from just four examples. The rules involved phonetic feature relations (same voicing or same place of articulation), and infants' learning was impeded when pairs of words allowed alternative possible generalizations (e.g. two words both contained the specific consonants p and t). Experiment 1 asked whether a small number of such spurious generalizations found in a randomly ordered list of 24 different words would also impede learning. It did - infants showed no sign of learning the rule. To ask whether it was the overall set of words or their order that prevented learning, Experiment 2 reordered the words to avoid local spurious generalizations. Infants showed robust learning. Infants thus appear to entertain spurious generalizations based on small, local subsets of stimuli. The results support a characterization of infants as incremental rather than batch learners.

You cannot speak and listen at the same time: a probabilistic model of turn-taking

Turn-taking is a preverbal skill whose mastering constitutes an important precondition for many social interactions and joint actions. However, the cognitive mechanisms supporting turn-taking abilities are still poorly understood. Here, we propose a computational analysis of turn-taking in terms of two general mechanisms supporting joint actions: action prediction (e.g., recognizing the interlocutor's message and predicting the end of turn) and signaling (e.g., modifying one's own speech to make it more predictable and discriminable). We test the hypothesis that in a simulated conversational scenario dyads using these two mechanisms can recognize the utterances of their co-actors faster, which in turn permits them to give and take turns more efficiently. Furthermore, we discuss how turn-taking dynamics depend on the fact that agents cannot simultaneously use their internal models for both action (or messages) prediction and production, as these have different requirements-or, in other words, they cannot speak and listen at the same time with the same level of accuracy. Our results provide a computational-level characterization of turn-taking in terms of cognitive mechanisms of action prediction and signaling that are shared across various interaction and joint action domains.

Learning about and from others' prudence, impatience or laziness: The computational bases of attitude alignment

Peoples' subjective attitude towards costs such as, e.g., risk, delay or effort are key determinants of inter-individual differences in goal-directed behaviour. Thus, the ability to learn about others' prudent, impatient or lazy attitudes is likely to be critical for social interactions. Conversely, how adaptive such attitudes are in a given environment is highly uncertain. Thus, the brain may be tuned to garner information about how such costs ought to be arbitrated. In particular, observing others' attitude may change one's uncertain belief about how to best behave in related difficult decision contexts. In turn, learning from others' attitudes is determined by one's ability to learn about others' attitudes. We first derive, from basic optimality principles, the computational properties of such a learning mechanism. In particular, we predict two apparent cognitive biases that would arise when individuals are learning about others’ attitudes: (i) people should overestimate the degree to which they resemble others (false-consensus bias), and (ii) they should align their own attitudes with others’ (social influence bias). We show how these two biases non-trivially interact with each other. We then validate these predictions experimentally by profiling people's attitudes both before and after guessing a series of cost-benefit arbitrages performed by calibrated artificial agents (which are impersonating human individuals).

What do people learn from observing others' attitudes, such as "prudence", "impatience" or "laziness"? Rather than viewing these attitudes as examples of highly subjective personality traits, we assume that they derive from uncertain (and mostly implicit) beliefs about how to best weigh risks, delays and efforts in ensuing cost-benefit trade-offs. In this view, it is adaptive to update one's belief after having observed others' attitude, which provides valuable information regarding how to best behave in related difficult decision contexts. This is the starting point of our computational model of attitude alignment, which we derive from first optimality principles as well as from recent neuroscientific findings. Critical here is the impact of one's ability to learn about others' covert mental states or attitudes, which is known as "mentalizing" or "Theory of Mind". In particular, this model makes two (otherwise unrelated) predictions that conform to known but puzzling cognitive biases of social cognition in humans, namely: "false consensus" and "social influence". It also shows how attitude alignment may eventually follow from the interaction between these two biases. Using state-of-the-art behavioural and computational methods, we provide experimental evidence that confirm these predictions. Finally, we discuss the relevance and implications of this work, both from a neuroscientific and economic perspective.

Social is special: A normative framework for teaching with and learning from evaluative feedback

Humans often attempt to influence one another's behavior using rewards and punishments. How does this work? Psychologists have often assumed that "evaluative feedback" influences behavior via standard learning mechanisms that learn from environmental contingencies. On this view, teaching with evaluative feedback involves leveraging learning systems designed to maximize an organism's positive outcomes. Yet, despite its parsimony, programs of research predicated on this assumption, such as ones in developmental psychology, animal behavior, and human-robot interaction, have had limited success. We offer an explanation by analyzing the logic of evaluative feedback and show that specialized learning mechanisms are uniquely favored in the case of evaluative feedback from a social partner. Specifically, evaluative feedback works best when it is treated as communicating information about the value of an action rather than as a form of reward to be maximized. This account suggests that human learning from evaluative feedback depends on inferences about communicative intent, goals and other mental states-much like learning from other sources, such as demonstration, observation and instruction. Because these abilities are especially developed in humans, the present account also explains why evaluative feedback is far more widespread in humans than non-human animals.

Building machines that learn and think like people

Recent progress in artificial intelligence has renewed interest in building systems that learn and think like people. Many advances have come from using deep neural networks trained end-to-end in tasks such as object recognition, video games, and board games, achieving performance that equals or even beats that of humans in some respects. Despite their biological inspiration and performance achievements, these systems differ from human intelligence in crucial ways. We review progress in cognitive science suggesting that truly human-like learning and thinking machines will have to reach beyond current engineering trends in both what they learn and how they learn it. Specifically, we argue that these machines should (1) build causal models of the world that support explanation and understanding, rather than merely solving pattern recognition problems; (2) ground learning in intuitive theories of physics and psychology to support and enrich the knowledge that is learned; and (3) harness compositionality and learning-to-learn to rapidly acquire and generalize knowledge to new tasks and situations. We suggest concrete challenges and promising routes toward these goals that can combine the strengths of recent neural network advances with more structured cognitive models.

Quantitative Pedagogy: A Digital Two Player Game to Examine Communicative Competence

Inner concepts are much richer than the words that describe them. Our general objective is to inquire what are the best procedures to communicate conceptual knowledge. We construct a simplified and controlled setup emulating important variables of pedagogy amenable to quantitative analysis. To this aim, we designed a game inspired in Chinese Whispers, to investigate which attributes of a description affect its capacity to faithfully convey an image. This is a two player game: an emitter and a receiver. The emitter was shown a simple geometric figure and was asked to describe it in words. He was informed that this description would be passed to the receiver who had to replicate the drawing from this description. We capitalized on vast data obtained from an android app to quantify the effect of different aspects of a description on communication precision. We show that descriptions more effectively communicate an image when they are coherent and when they are procedural. Instead, the creativity, the use of metaphors and the use of mathematical concepts do not affect its fidelity.

A Duet for one

This paper considers communication in terms of inference about the behaviour of others (and our own behaviour). It is based on the premise that our sensations are largely generated by other agents like ourselves. This means, we are trying to infer how our sensations are caused by others, while they are trying to infer our behaviour: for example, in the dialogue between two speakers. We suggest that the infinite regress induced by modelling another agent - who is modelling you - can be finessed if you both possess the same model. In other words, the sensations caused by others and oneself are generated by the same process. This leads to a view of communication based upon a narrative that is shared by agents who are exchanging sensory signals. Crucially, this narrative transcends agency - and simply involves intermittently attending to and attenuating sensory input. Attending to sensations enables the shared narrative to predict the sensations generated by another (i.e. to listen), while attenuating sensory input enables one to articulate the narrative (i.e. to speak). This produces a reciprocal exchange of sensory signals that, formally, induces a generalised synchrony between internal (neuronal) brain states generating predictions in both agents. We develop the arguments behind this perspective, using an active (Bayesian) inference framework and offer some simulations (of birdsong) as proof of principle.

Faster Teaching via POMDP Planning

Human and automated tutors attempt to choose pedagogical activities that will maximize student learning, informed by their estimates of the student's current knowledge. There has been substantial research on tracking and modeling student learning, but significantly less attention on how to plan teaching actions and how the assumed student model impacts the resulting plans. We frame the problem of optimally selecting teaching actions using a decision-theoretic approach and show how to formulate teaching as a partially observable Markov decision process planning problem. This framework makes it possible to explore how different assumptions about student learning and behavior should affect the selection of teaching actions. We consider how to apply this framework to concept learning problems, and we present approximate methods for finding optimal teaching actions, given the large state and action spaces that arise in teaching. Through simulations and behavioral experiments, we explore the consequences of choosing teacher actions under different assumed student models. In two concept-learning tasks, we show that this technique can accelerate learning relative to baseline performance.

Young Children's Help-Seeking as Active Information Gathering

Young children's social learning is a topic of great interest. Here, we examined preschoolers' (M = 52.44 months, SD = 9.7 months) help-seeking as a social information gathering activity that may optimize and support children's opportunities for learning. In a toy assembly task, we assessed each child's competency at assembling toys and the difficulty of each step of the task. We hypothesized that children's help-seeking would be a function of both initial competency and task difficulty. The results confirmed this prediction; all children were more likely to seek assistance on difficult steps and less competent children sought assistance more often. Moreover, the magnitude of the help-seeking requests (from asking for verbal confirmation to asking the adult to take over the task) similarly related to both competency and difficulty. The results provide support for viewing children's help-seeking as an information gathering activity, indicating that preschoolers flexibly adjust the level and amount of assistance to optimize their opportunities for learning.

Verbal framing of statistical evidence drives children's preference inferences

Although research has shown that statistical information can support children's inferences about specific psychological causes of others' behavior, previous work leaves open the question of how children interpret statistical information in more ambiguous situations. The current studies investigated the effect of specific verbal framing information on children's ability to infer mental states from statistical regularities in behavior. We found that preschool children inferred others' preferences from their statistically non-random choices only when they were provided with verbal information placing the person's behavior in a specifically preference-related context, not when the behavior was presented in a non-mentalistic action context or an intentional choice context. Furthermore, verbal framing information showed some evidence of supporting children's mental state inferences even from more ambiguous statistical data. These results highlight the role that specific, relevant framing information can play in supporting children's ability to derive novel insights from statistical information.

The social Bayesian brain: does mentalizing make a difference when we learn?

When it comes to interpreting others' behaviour, we almost irrepressibly engage in the attribution of mental states (beliefs, emotions…). Such "mentalizing" can become very sophisticated, eventually endowing us with highly adaptive skills such as convincing, teaching or deceiving. Here, sophistication can be captured in terms of the depth of our recursive beliefs, as in "I think that you think that I think…" In this work, we test whether such sophisticated recursive beliefs subtend learning in the context of social interaction. We asked participants to play repeated games against artificial (Bayesian) mentalizing agents, which differ in their sophistication. Critically, we made people believe either that they were playing against each other, or that they were gambling like in a casino. Although both framings are similarly deceiving, participants win against the artificial (sophisticated) mentalizing agents in the social framing of the task, and lose in the non-social framing. Moreover, we find that participants' choice sequences are best explained by sophisticated mentalizing Bayesian learning models only in the social framing. This study is the first demonstration of the added-value of mentalizing on learning in the context of repeated social interactions. Importantly, our results show that we would not be able to decipher intentional behaviour without a priori attributing mental states to others.