Special Journal Issue on Interfacing Mind and Environment: The Central Role of Search in Cognition in Topics in Cognitive Science
Wai-Tat Fu (University of Illinois at Urbana-Champaign),
Thomas Hills (Warwick University), &
Peter Todd (Indiana University)
Interfacing Mind and Environment: The Central Role of Search in Cognition
Search can be found in almost every cognitive activity, ranging across vision, memory retrieval, problem solving, decision making, foraging, and social interaction. Because of its ubiquity, research on search has had a tendency to fragment into multiple areas of cognitive science. The proposed topic aims at providing an integrative discussion of the central role of search from multiple perspectives. We focus on controlled search processes, which require (1) a goal, (2) uncertainty about the nature, location, or acquisition method of the objects to be searched for, and (3) a method for sampling through the search environment. While this definition of search is general and applicable to different domains, the specific search environments, strategies, and underlying cognitive and neural processes may differ. The goal of this issue is to compare and contrast search processes, in an effort to understand how structure, strategy, and process interact to generate search across different cognitive domains. We expect that given its cross-domain nature, the topic on search will be of broad interest to cognitive scientists including psychologists, behavioral ecologists, computer scientists, neuroscientists, linguists, and sociologists.
We have invited 8 target articles to be published in this issue of Topics in Cognitive Science. We are interested in soliciting commentaries for each of these target articles. The goal is to foster more general and deeper understanding of cognitive search. We will announce details of how we will solicit commentaries soon. The titles, authors, and abstracts of the target articles can be found below.
Integration of social information by human groups
Boris Granovskiy1, Jason M. Gold2, David Sumpter1, Robert L. Goldstone2
1Uppsala University, 2Indiana University
We consider a situation in which individuals search for accurate decisions without direct feedback on their accuracy but with information about the decisions made by peers in their group. The “wisdom of crowds” hypothesis states that the average judgment of many individuals can give a good estimate of, for example, the outcomes of sporting events and the answers to trivia questions. Two limitations of the application of wisdom of crowds are that estimates should be independent and unbiased. Here, we study how individuals integrate social information when answering trivia questions with answers between 0 and 100%. We find that, consistent with wise crowds, average performance improves with group size. However, individuals show a consistent bias to produce estimates that are insufficiently extreme. We find that social information provides significant, albeit small, improvement in group performance. Outliers with answers far from the correct answer move towards the position of the group mean. Given that these outliers also tend to be nearer to 50% than do the answers of other group members, this move creates group polarization away from 50%. By looking at individual performance over different questions we find that some people are more susceptible to social influence than others. There is also some evidence that people differ in their competence in answering questions, but lack of competence is not significantly correlated with willingness to change guesses. We develop a mathematical model based on these results to show that communication can serve to remove outlying incorrect opinions and, in some situations, improve group performance. However, improvement is only predicted for cases in which the initial guesses of individuals in the group are biased.
Novelty and Inductive Generalization in Human Reinforcement Learning
Samuel J. Gershman and Yael Niv (Princeton University)
In reinforcement learning, a decision maker searching for the most rewarding option is often faced with the question: what is the value of an option that has never been tried before? One way to frame this question is as an inductive problem: how can I generalize my previous experience with one set of options to a novel option? We show how hierarchical Bayesian inference can be used to solve this problem, and describe an equivalence between the Bayesian model and temporal difference learning algorithms that have been proposed as models of reinforcement learning in humans and animals. According to our view, the search for the best option is guided by abstract knowledge about the relationships between different options in an environment, resulting in greater search efficiency compared to traditional reinforcement learning algorithms previously applied to human cognition. In three behavioral experiments, we test several predictions of our model, providing evidence that humans learn and exploit structured inductive knowledge to make predictions about novel options. In light of this model, we suggest a new interpretation of dopaminergic responses to novelty.
Optimal Semantic Search in the Remote Associates Test
Eddy Davelaar (University of London, Birkbeck)
Searching through semantic memory may involve the use of several retrieval cues. In a verbal fluency task, the set of available cues is limited and every candidate word is a target. Individuals exhibit clustering behaviour as predicted by optimal foraging theory. In another semantic search task, the remote associates task (RAT), three cues are presented and a single target word has to be found. Whereas the task has been widely studied as a task of creativity or insight problem solving, in this article, the RAT is treated as a semantic retrieval task and assessed from the perspective of information foraging theory. Experiments are presented that address the superadditive combination of cues and the anti-clustering behaviour in the recall sequence. A consistent information accumulation hypothesis of search behaviour in the RAT is put forward in which optimal search in RAT problems involves maximising the difference in activation between target and distractors. This type of search is optimal when the target is weak and cue patches are contaminated with strong competitors.
Search and the Aging Mind: The Promise and Limits of the Cognitive Control Hypothesis of Age Differences in Search
Rui Mata and Bettina von Helversen (University of Basel)
Search is a prerequisite for successful performance in a broad range of tasks ranging from making decisions between consumer goods to memory retrieval. How does aging impact search processes in such disparate situations? Aging is associated with structural and neuromodulatory brain changes that underlie cognitive control processes, which in turn have been proposed as a domain-general mechanism underlying search. We review the aging literature to evaluate the cognitive control hypothesis that suggests that age-related change in cognitive control underlies age differences in both external and internal search. We also consider the limits of the cognitive control hypothesis and propose additional mechanisms such as changes in strategy use and affect that may be necessary to understand how aging affects search.
Environment structure, Motivation and Search Behavior
Arthur B. Markman, A. Ross Otto, Brian Glass, W. Todd Maddox, & Bradley C. Love (University of Texas at Austin)
Many decision making situations require trading off between exploring a space of choices and exploiting knowledge of past choices. The degree to which people prefer to explore or exploit can be influenced by a number of different factors. In this paper, we focus on the role of the structure of the environment and the influence of motivational factors on exploration. These issues are examined in studies of foraging in two-dimensional environments and in repeated choices in N-arm bandit problems. In each domain, we support the results with simulations of people's performance that helps to clarify the relationship between their performance and underlying cognitive variables.
Individual differences in animal exploration and search: Social and developmental influences
Simon Reader (McGill University)
Numerous studies have documented individual differences in exploratory tendencies and other phenomena related to search, and these differences have been linked to fitness. Here, I discuss the origins of these differences, focusing on how experience shapes animal search and exploration. The origin of individual differences will also depend upon the alternatives to exploration that are available. Given that search and exploration frequently carry significant costs, we might expect individuals to utilize cues indicating the potential net payoffs of exploration versus the exploitation of known acts. Informative cues could arise from both recent and early-life experiences, from both the social and physical environment. Open questions are the extent to which individual search tendencies are fixed throughout life versus being flexibly adjusted according to prevailing conditions and the actions of other individuals, and the extent to which individual differences in exploration extend across domains and are independent of other processes.
Information Foraging across the Life Span: Search and Switch in Unknown Patches
Jessie Chin, Brennan Payne, Andrew Battles, Wai-Tat Fu, Dan Morrow and Elizabeth A. L. Stine-Morrow (University of Illinois at Urbana-Champaign)
The study examined the effects of task difficulty, patch heterogeneity and age on individual differences in information uptake and switch behavior in an information foraging task. 30 younger and 30 older participants had limited time to find words in a set of 4 word search puzzles on an electronic tablet that recorded performance, search time and switches between puzzles. There were three conditions: all easy, all puzzles containing high‐prototypical category exemplars in canonical orientations in the puzzle (forward, down); all difficult, puzzles containing low‐prototypical exemplars in any orientations (e.g., forward and backward diagonal); and mixed (2 easy, 2 difficult puzzles). Mixed effects modeling was used to estimate the rates of information gain (RIGs); i.e., cumulative number of words found as a function of time, for each participant. RIGs varied as a function of difficulty, bouts (i.e., attempts) and age. For example, younger adults had similar RIGs for difficult puzzles across bouts, but older adults showed a decrease in RIGs across bouts. At the same time, older adults persisted longer in the most difficult condition, which was adaptive in enabling them to achieve information uptake comparable to the younger adults. The revisit to the puzzles was especially adaptive for the hard puzzles and participants who switched more often. Older adults switched among information patches less often than younger ones suggesting that older adults tended to persist in the patch relatively longer when RIGs were dropping. Overall, the study suggests that age‐related strategic variation in information foraging is shaped by both individual capacity and environmental context.
Semantic Fields Forever: Defining the Patch in Memory Search
Thomas Hills1, Peter Todd2, and Michael Jones2
(1Warwick University, 2Indiana University)
When searching for concepts of some type in memory—as in the verbal fluency task of naming all the tools one can think of—people appear to explore internal mental representations in much the same way that animals forage in physical space—searching locally within patches of information before transitioning globally between patches. However, the definition of a patch in mental space is not well specified. Do people activate categories when searching mental space (categorical search), or do they activate particular items and then search for nearby items (associative search), or both? Using semantic representations in a Search of Associative Memory (SAM) framework, we tested alternative hypotheses based on associative and categorical patch models. The results are generally consistent with an associative patch model, but some individuals appear to use a categorical patch model. However, we found no relationship between best fitting patch models and performance. In sum, the question remains open and may represent a strategic difference in the way internal information is searched or represented by different individuals.