Patrick Shafto

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 students 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.

Controlling the message: preschoolers’ use of information to teach and deceive others

Effective communication entails the strategic presentation of information; good communicators present representative information to their listeners—information that is both consistent with the concept being communicated and also unlikely to support another concept a listener might consider. The present study examined whether preschool-age children effectively select information to manipulate others’ semantic knowledge, by testing how children choose information to teach or deceive their listeners. Results indicate that preschoolers indeed effectively select information to meet some specific communicative goals. When asked to teach others, children selected information that effectively spanned the concept of interest and avoided overly restrictive or overly general information; when asked to deceive others, they selected information consistent with the intended deceptive messages under some circumstances. Thus, preschool children possess remarkable abilities to select the best information to manipulate what others believe.

Human-Recommender Systems: From Benchmark Data to Benchmark Cognitive Models

We bring to the fore of the recommender system research community, an inconvenient truth about the current state of understanding how recommender system algorithms and humans influence one another, both computationally and cognitively. Unlike the great variety of supervised machine learning algorithms which traditionally rely on expert input labels and are typically used for decision making by an expert, recommender systems specifically rely on data input from non-expert or casual users and are meant to be used directly by these same non-expert users on an every day basis. Furthermore, the advances in online machine learning, data generation, and predictive model learning have become increasingly interdependent, such that each one feeds on the other in an iterative cycle. Research in psychology suggests that peoples choices are (1) contextually dependent, and (2) dependent on interaction history. Thus, while standard methods of training and assessing performance of recommender systems rely on benchmark datasets, we suggest that a critical step in the evolution of recommender systems is the development of benchmark models of human behavior that capture contextual and dynamic aspects of human behavior. It is important to emphasize that even extensive real life user-tests may not be sufficient to make up for this gap in benchmarking validity because user tests are typically done with either a focus on user satisfaction or engagement (clicks, sales, likes, etc) with whatever the recommender algorithm suggests to the user, and thus ignore the human cognitive aspect. We conclude by highlighting the interdisciplinary implications of this endeavor.

Epistemic Trust and Education: Effects of Informant Reliability on Student Learning of Decimal Concepts.

The epistemic trust literature emphasizes that childrens evaluations of informants trustworthiness affects learning, but there is no evidence that epistemic trust affects learning in academic domains. The current study investigated how reliability affects decimal learning. Fourth and fifth graders (N = 122; Mage = 10.1 years) compared examples from consistently accurate and inaccurate informants (consistent) or informants who were each sometimes accurate and inaccurate (inconsistent). Fourth graders had higher conceptual knowledge and fewer misconceptions in the consistent condition than the inconsistent condition, and vice versa for fifth graders due to differences in prior exposure to decimals. Given the same examples, learning differed depending on informant reliability. Thus, epistemic trust is a malleable factor that affects learning in an academic domain.

Questioning supports effective transmission of knowledge and increased exploratory learning in pre-kindergarten children

How can education optimize transmission of knowledge while also fostering further learning? Focusing on children at the cusp of formal schooling (N = 180, age = 4.0-6.0 y), we investigate learning after direct instruction by a knowledgeable teacher, after questioning by a knowledgeable teacher, and after questioning by a naïve informant. Consistent with previous findings, instruction by a knowledgeable teacher allows effective information transmission but at the cost of exploration and further learning.xa0Critically, we find a dual benefit for questioning by a knowledgeable teacher: Such pedagogical questioning both effectively transmits knowledge and fosters exploration and further learning, regardless of whether the question was directed to the child or directed to a third party and overheard by the child. These effects are not observed when the same question is asked by a naïve informant. We conclude that a teachers choice of pedagogical method may differentially influence learning through their choices of how, and how not, to present evidence, with implications for transmission of knowledge and self-directed discovery. A video abstract of this article can be viewed at: https://www.youtube.com/watch?v=FJXH2b65wL8.

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.

Pedagogical Questions in Parent–Child Conversations

Questioning is a core component of formal pedagogy. Parents commonly question children, but do they use questions to teach? This article defines pedagogical questions as questions for which the questioner already knows the answer and intended to help the questionee learn. Transcripts of parent-child conversations were collected from the CHILDES database to examine the frequency and distribution of pedagogical questions. Analysis of 2,166 questions from 166 mother-child dyads and 64 father-child dyads (childs age between 2 and 6xa0years) showed that pedagogical questions are commonplace during day-to-day parent-child conversations and vary based on childs age, family environment, and historical era. The results serve as a first step toward understanding the role of parent-child questions in facilitating childrens learning.

Development of children’s sensitivity to overinformativeness in learning and teaching.

Effective communication requires knowing the “right” amount of information to provide; what is necessary for a naïve learner to arrive at a target hypothesis may be superfluous and inefficient for a knowledgeable learner. The current study examines 4- to 7-year-olds’ developing sensitivity to overinformative communication and their ability to decide how much information is appropriate depending on the learner’s prior knowledge. In Experiment 1 (N = 184, age = 4.09–7.98 years), 5- to 7-year-old children preferred teachers who gave costly, exhaustive demonstrations when learners were naïve, but preferred teachers who gave efficient, selective demonstrations when learners were already knowledgeable given their prior experience (i.e., common ground). However, 4-year-olds did not show a clear preference. In Experiment 2 (N = 80, age = 4.05–6.99 years), we asked whether children flexibly modulated their own teaching based on learners’ knowledge. Five and 6-year-olds, but not 4-year-olds, were more likely to provide exhaustive demonstrations to naïve learners than to knowledgeable learners. These results suggest that by 5 years of age, children are sensitive to overinformativeness and understand the trade-off between informativeness and efficiency; they reason about what others know based on the presence or absence of common ground and flexibly decide how much information is appropriate both as learners and as teachers.

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.

The relationship between non-symbolic multiplication and division in childhood

Children without formal education in addition and subtraction are able to perform multi-step operations over an approximate number of objects. Further, their performance improves when solving approximate (but not exact) addition and subtraction problems that allow for inversion as a shortcut (e.g., a + b − b = a). The current study examines childrens ability to perform multi-step operations, and the potential for an inversion benefit, for the operations of approximate, non-symbolic multiplication and division. Children were trained to compute a multiplication and division scaling factor (*2 or /2, *4 or /4), and were then tested on problems that combined two of these factors in a way that either allowed for an inversion shortcut (e.g., 8*4/4) or did not (e.g., 8*4/2). Childrens performance was significantly better than chance for all scaling factors during training, and they successfully computed the outcomes of the multi-step testing problems. They did not exhibit a performance benefit for problems with the a*b/b structure, suggesting that they did not draw upon inversion reasoning as a logical shortcut to help them solve the multi-step test problems.