- About Us
- Affiliates and Liaisons
Colloquia occur: Selected Mondays at 4:00 pm - 5:00 pm - Room PY 101.
Colloquia titles will be posted as they become available.
Organizer: Rob Goldstone
Sep 9, 2013: Morana Alac (UCSD)
Title: Being social as rooted in practical situations of interaction
Abstract: This talk will report on a long-term and ongoing observational study of activities around a research project in social robotics. The roboticists whose work I observe design educational robots geared toward children between 18 and 24 months of age. To do so, they work on the computational architecture, while continually updating the robot’s current version by immersing it in a preschool setting. In this talk I will neither inform on the robot’s computational architecture nor provide suggestions for the future improvements of the robot's design. Instead, my interest is in social agency and how it is instantiated at the preschool. That social robots provide an opportunity to specify what it means to be an agent in design terms is one of their major appeals. I will show videotaped excerpts from everyday interactions at the preschool to suggest a somewhat different take. Grounded in local, embodied practices that those who engage with the robot and each other recognize and understand, I will discuss the robot’s agency as a local, multiparty achievement that is sensitive to situational particularities and interactional contingencies.
Sep 23, 2013: Chris Eliasmith (University of Waterloo)
Title: How to Build a Brain
Abstract: Recent high-profile brain simulations, including those of the $1 billion Euro Human Brain Project, are very large and complex. However, they do not exhibit interesting behaviours and so are difficult to compare to much of what we know about the brain. In this talk, I describe the methods and tools used to construct what is currently the world's largest *functional* brain simulation, called the Semantic Pointer Architecture Unified Network (Spaun). I demonstrate the variety of behaviors the model exhibits and show that it is similar in many respects to human and animal behaviour. I argue that constructing such large-scale simulations is critical for advancing our understanding of neural and cognitive function.
Sep 30, 2013: Alison Preston (University of Texas)
Title: Building knowledge through memory integration
Abstract: New learning experiences are not isolated from the past. Our ability to learn new information is often influenced by what we already know. The idea that new events in the external world are processed in reference to pre-existing memories is characteristic of several influential psychological theories; yet, virtually nothing is known about how the human brain uses prior experience to make sense of new events. In this talk, I will discuss our recent human neuroimaging work characterizing the neural mechanisms that support the integration of new knowledge into existing memory frameworks. In a series of studies, we show that the hippocampus and ventromedial prefrontal cortex play key roles in constructing integrated memory networks and further demonstrate that these integrated memories are used in service of future decisions. Collectively, our data provide novel evidence for the view that memories do not simply consist of individual records of directly experienced events, but also include memories built by integrating knowledge across distinct experiences. In doing so, our findings emphasize the adaptive nature of memory, whereby memory representations are constructed to anticipate and successfully negotiate future judgments.
Oct 14, 2013: Jesse Snedecker (Harvard University)
Title: Children's language comprehension: incremental, interactive and abstract
Abstract: In the last thirty years, cognitive science has converged on a model of human information processing with three characteristics\: levels of representation, the incremental propagation of information from one level to the next, and interactivity \(processes at one level are influenced by information at multiple other levels\). My research uses the visual world paradigm and priming techniques to demonstrate that children's moment-to-moment language comprehension has all three of these properties. These tools can be used to explore the nature of the representations that children construct during language processing, their ability to make pragmatic inferences, and how these processes break down in developmental disorders.
Oct 21, 2013: Bill Warren (Brown University)
Title: Self-organization in human crowds: From individual to collective behavior
Abstract: The collective behavior of human crowds is thought to emerge from local interactions between pedestrians. There are many agent-based models of such “flocking” behavior in animals and humans, but precious little experimental data. I will describe an empirically-grounded model of pedestrian and crowd dynamics, following what Sumpter, et al. (2012) call the modeling cycle. First, taking a local-to-global approach, we use experiments in VR to create a model of individual locomotor behavior and neighbor interactions. This pedestrian model is then implemented in multi-agent simulations to test whether it can generate global patterns of crowd behavior. Conversely, taking a global-to-local approach, we collect motion-capture data on human crowds (N=20) to test the simulations. Patterns of crowd behavior are analyzed to estimate the local coupling between neighbors and refine the model. I will report simulations of several crowd scenarios, including (a) Grand Central, in which individuals criss-cross the room, (b) swarm, in which participants walk about while staying together as a group, and (c) counterflow, in which two groups pass through each other. Each global pattern can be simulated with a few model components. Analysis of the swarm data suggests a strong but fairly local coupling (1-2 m). The results support the view that crowd dynamics emerge from a few simple pedestrian interactions, and may be accounted for by principles of self-organization.
Oct 28, 2013: Steven Sloman (Brown University)
Title: Knowledge and Illusion in Causal Reasoning: The Right and Wrong of Bayes Nets
Abstract: A promising normative framework for causal reasoning has become popular in the form of Causal Bayesian Networks. I report some empirical tests of qualitative properties of the framework. The framework has been challenged by violations of its key representational assumption (the Markov property or screening-off). The assumption can be defended once people's causal models are fleshed out with their mechanistic assumptions. I also show some limitations of the framework for modeling human judgment. First, the framework expects more symmetry between predictive and diagnostic reasoning than people show. For instance, people tend to neglect alternative causes when reasoning predictively but not diagnostically. A consequence is that people treat weak positive evidence as negative. Second, people typically know less about causal system than they think they do (the illusion of explanatory depth). I report studies showing this is true in politics as well and that shattering the illusion leads to more moderate attitudes and can reduce donations to relevant political advocacy groups. The illusion is not present in those who score well on the Cognitive Reflection Test.
Nov 11, 2013: Bethany Rittle-Johnson (Vanderbilt University)
Title: The Power of Comparison in Learning and Instruction: Experimental Evidence from Mathematics Classsrooms
Abstract: Comparison is a powerful learning process that has been leveraged to improve learning in many domains. I will synthesize results from my research on comparison with Jon Star. In 5 short-term experimental, classroom-based studies, we found that comparing multiple methods for solving the same problem consistently supported procedural flexibility, and sometimes supported greater conceptual and procedural knowledge. We next developed a supplemental Algebra 1 curriculum that fosters comparison and conducted a randomized-control trial with 76 Algebra I teachers. Findings were more modest in this study. I will illustrate how cognitive science research helped guide the design of effective educational materials and how educational practice revealed new directions for future research.Comparison is a powerful learning process that has been leveraged to improve learning in many domains. I will synthesize results from my research on comparison with Jon Star. In 5 short-term experimental, classroom-based studies, we found that comparing multiple methods for solving the same problem consistently supported procedural flexibility, and sometimes supported greater conceptual and procedural knowledge. We next developed a supplemental Algebra 1 curriculum that fosters comparison and conducted a randomized-control trial with 76 Algebra I teachers. Findings were more modest in this study. I will illustrate how cognitive science research helped guide the design of effective educational materials and how educational practice revealed new directions for future research.
Dec 2, 2013: Nick Turk-Browne (Princeton University)
Title: Statistical learning in the mind and brain
Abstract: The environment is highly stable over time. Not only do we repeatedly encounter the same people, places, and objects, but they also tend to appear in regular sequences and configurations. These regularities are extracted rapidly and without effort or awareness, resulting in higher-order knowledge of words, events, and scenes. With behavioral studies, I will suggest that this kind of statistical learning has widespread consequences for the mind, including showing that it commands attention to locations and features with structure. With neuroimaging and patient work, I will then suggest that the medial temporal lobe memory system, and hippocampus in particular, implements this kind of statistical learning in the brain, including showing that this region learns by shaping the representational space for objects to mirror the structure of the environment. These studies reveal the ubiquity and power of statistical learning, and they highlight the inherently reciprocal connections between perception, attention, and memory.
Dec 9, 2013: Daphne Maurer (McMaster University)
Title: Critical Periods Re-examined: Evidence from Children treated for Dense Cataracts
Abstract: We have been taking advantage of a natural experiment: children treated for dense cataracts that blocked all patterned vision to the retina until the cataracts were removed surgically and the eyes fit with compensatory contact lenses. I will describe the general principles that have emerged from comparing the effects of bilateral and unilateral cataracts and from studying the consequences of deprivation that began at different ages. Together, the results suggest different critical periods for damaging different aspects of vision and different principles for low level (e.g., acuity) and higher level vision (e.g., global motion; face processing). Nevertheless, some potential for rehabilitation remains even in adulthood.