CogSci 2018: Madison
July 25 – July 28th, 2018
The conference workshop program provides an opportunity for in-depth discussion on a specific topic important to cognitive science. The workshops will be held the day before the main conference, on July 25, 2018. The location will be the same as the main conference.
Learning as Program Induction
Neil Bramley, Eric Schulz, Fei Xu, Josh Tenenbaum
This workshop will cover new work that casts human learning as program induction — i.e. learning of programs from data. This workshop will bring together scientists who have a joint interest in how intelligent systems (humans or machines) can learn rich representations and action plans (expressable as programs) though observing and interacting with the world. It will involve talks by a number of leading researchers from cognitive science, computational linguistics, and artificial intelligence interested in the computational foundations of reasoning and round off with an open discussion session.
Conceptual foundations in dynamic field theory: Applications in cognitive and developmental science
Aaron Buss, Sammy Perone, Ajaz Ahmad Bhat
Dynamical systems thinking has been influential in the ways psychologists, cognitive scientists, and neuroscientists think about cognition and development. Dynamic field theory (DFT) has applied the concept of dynamic activation fields to address fundamental questions relating to cognition. DFT provides important tools for linking cognition with sensori-motor processes, linking timescales from in-the-moment behavior to learning and development, and linking levels of analysis between neural and behavioral data. One obstacle for researchers wishing to use DFT has been that the foundational concepts of DFT are distinct from the standard way of thinking about cognition as comprised of static components. The computational foundations of DFT are also distinct from many traditional computational frameworks that are popular in the fields of cognitive and developmental sciences. Thus, the goal of this tutorial is to provide the training and tools to overcome these obstacles. We will provide a systematic introduction to the central concepts of DFT and their grounding in both dynamical systems concepts and neurophysiology. We will discuss the concrete mathematical implementation of these concepts in dynamic neural field models, giving all needed background information. Additionally, participants will have the opportunity to gain hands-on experience using interactive simulators in MATLAB. Finally, we will take participants through selected, exemplary case studies in which the concepts and associated models have been used to ask questions about elementary forms of embodied cognition and their development.
Understanding Exploration-Exploitation Trade offs
Elizabeth Bonawitz, Alison Gopnik, Celeste Kidd
All cognitive systems both gather information and use that information to make decisions (e.g., explore and exploit). But there is an intrinsic tension between these two aspects of cognition. Exploration is costly in that resources allocated to information-seeking are unavailable for action. Moreover, actions have to be taken before all relevant information is available. When should a cognitive system seek new information and when should it stop to decide how to act? How do people and other organisms resolve this tension between exploration and exploitation? Is there an optimal way of doing so? And how can systems be motivated to seek information when that pursuit is costly and the immediate utility of the information may not be apparent? The workshop will feature well-known experts from different fields (more information at this website:). We are also inviting poster submissions**, with “poster teasers” flash talks and an opportunity to view new studies investigating exploration-exploitation. **Submission deadline is June16th; apply here: https://sites.google.com/site/cogsci2018exploreexploit/poster-submission
Massive Online Experiments in Cognitive Science
Josh de Leeuw, Joshua K. Hartshorne, Laura Germine, Katharina Reinecke, Mariela V. Jennings
This full-day workshop focuses on Massive Online Experiments (MOEs). MOEs have transformative potential, as they effectively allow researchers to run hundreds of experiments simultaneously. The goal of the workshop is to help a broad cross-section of cognitive scientists begin to incorporate MOEs into their research. The morning session focuses on the methodology of MOEs and examples of research using MOEs. The afternoon is a practical introduction to tools for conducting MOEs. Attendees who have conducted an MOE are invited to present their work during the morning session. Depending on the interest level, this presentation will either be part of a series of rapid oral presentations or a poster session. If you are interested in presenting your work, please contact Josh de Leeuw at email@example.com with a title and short description of the project by July 1st.
Computational Methods and Systems for the Cognitive Modelling and Support of Creativity and Creative Problem Solving
Computational creativity and human creativity are fields modelled with different processes, and evaluated with different methods. To bridge this interdisciplinary divide, we need to (i) disseminate and refine existing computational methods for modelling cognitive processes, (ii) aim to implement more cognitive processes in computational creativity systems and (iii) set benchmarks of comparative evaluation between cognitive human and computational systems. Various computational methods might lend themselves better to modelling cognitive processes – for example semantic networks might help model associativity processes, case base reasoning might help model cognitive structured representations which admit variations, etc.
In this workshop we will discuss existing computational methods, systems and models, focusing on questions like the following: (i) what computational methods are more suitable for implementing computational models of creativity and problem solving, and computational systems supporting creativity and problem solving; (ii) what types of support can natural cognitive systems benefit from when performing creative problem solving and other creative acts; (iii) what kind of computational support has been offered so far, what kind of computational support can be offered with the existing techniques and approaches; (iv) to what extent computational methods must get closer to simulating or modeling cognitive process to make cognitive support possible.
The workshop will involve three elements. Invited speakers will present existing methods and system. Short presentations of papers and posters will be accepted on the topic. The workshop will end with a panel discussion, focused on establishing future directions for methods and systems aimed at supporting creativity and problem solving.
Contemporary Cognitive Approaches to Decision-Making
Oleg Urminsky, Daniel Bartels, Todd M Gureckis, Jennifer Trueblood
The study of how people make judgments and decisions began in cognitive science (the primary JDM conference literally was founded as a workshop at the Psychonomics conference). However, over time the area of judgment and decision-making (JDM) has moved apart from its cognitive roots, despite the high overlap in underlying research questions. However, greater interaction between cognitive and JDM research could yield benefits to both from cross-pollination (Bartels and Johnson 2015), including in terms of methods, types of relevant data and underlying questions. This workshop is designed to foster such interaction. The talks will explore productive areas of overlap between cognitive research and judgment and decision-making research.