2017 Fiscal Year Research-status Report
An adjoint functors approach to models of cognition
| Project/Area Number |
16KT0025
|
|---|
| Research Institution | National Institute of Advanced Industrial Science and Technology |
|---|
Principal Investigator |
Phillips Steven 国立研究開発法人産業技術総合研究所, 人間情報研究部門, 上級主任研究員 (90344209)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
武田 裕司 国立研究開発法人産業技術総合研究所, 自動車ヒューマンファクタ研究センタ, 研究チーム長 (10357410)
|
|---|
| Project Period (FY) |
2016-07-19 – 2019-03-31
|
| Keywords | dual-process / visual search / conjunctive search / least-cost / category theory / functor / adjunction / cognition |
|---|
| Outline of Annual Research Achievements |
Dual-process theories of cognition suppose that there are two styles of thinking: Type 1, which is fast, reflexive and relatively effortless, and Type 2, which is slow, reflective and relatively effortful. Although such theories can provide better fits to the data, they fail to provide satisfactory answers to two important questions. (1) What principle governs the invocation of one cognitive style over the other? (2) What is the relationship between Type 1 and Type 2 processes?
In regard to the first question, we proposed a least-cost hypothesis: participants select the process style of least cost. To test this hypothesis, we developed a split-screen paradigm, where participants could make a correct response by attending to stimuli presented in either the left or right field of view. The results support a contextualized form of the least-cost hypotheses: participants selected the alternative of least-cost, but only when the cost of assessing the alternatives was relatively low (Phillips et al, 2017).
In regard to the second question, we theorized that Type 1 and Type 2 processes are related by a category theory construction, called an adjunction: a pair of functors are pseudo-inverses of each other. Many common characterized distinctions between Type 1 and Type 2 derive from the properties of specific pairs of adjoint functors, thus lending support to the theory of adjoint functors as a basis for dual-process cognition (Phillips, 2017; to appear).
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
Experiments were conducted to test the least-cost hypothesis. Participants (23 adults) performed a visual search task: locate a target object in the left or right display field. Displays were arranged so that the target was identifiable by a single feature dimension, e.g., colour (feature search), or multiple features dimensions, e.g., colour and shape (conjunction search), at two levels of difficulty (easy, hard), by manipulating feature discriminability, and number of feature conjunctions (Fig. 1). Response times, error rates, and gaze rates were recorded.
|
| Strategy for Future Research Activity |
The experiments in 2017 showed that subjects were sensitive to relative task difficulty. However, one could argue that the alternative search routes were elicited the same basic (search) process under different input conditions, rather than two distinct types cognitive process. Hence, the plan is to test this situation using a visual marking version of our split-screen paradigm. In marked search, some of the search items are displayed prior to displaying the entire search field. It is well-known that subjects flag these items, so as exclude them from subsequent (re)search to yield more efficient search. Thus, visual search using prior marked items is known to involve cognitive processes that differ from visual search without marking. So we propose testing subjects on marked versus unmarked visual search version of our split-screen paradigm. That is, one field contains marking, the other field does not. Although marked search is generally more efficient, we can independently vary cost by including a delay between onset of mark versus unmark search fields. Our theory predicts that subjects should search of the lower cost side relative to control (one-sided version), rather than simply search on the marked side regardless of cost. We also plan to develop a computational model of the processes that govern route (process) selection.
|
| Causes of Carryover |
The budget for the new financial year is intended to cover the cost of:conducting new experiments,employing a technical assistant to manage administrative aspects of the project, and a technical assistant to implement computational models of the data generated from project experiments, and expenses associated with presenting results for conferences and journal publications.
|
