Age-Related Differences in Object-Similarity Judgment

• Author / Creator

  McKnight, D E.

• As people age, they adjust how they associate objects. We examine the interplay between age and various aspects of object similarity. For this, we perform two experiments: one between adults aged 25–35 and 50–60 and another between six-year-olds and adults. Between 25–35- and 50–60-year-olds, we investigate discrepancies in preference for each of 49 object-comparison dimensions. Between six-year-olds and adults, we investigate changes in the prioritization of the classes of object-comparison relations known as taxonomic and thematic relations. To facilitate these tasks, we use a prior interpretable, machine-learned computational model; this model is trained to perform an odd-one-out-among-three task with a vector embedding for each object being compared.

  For the first task, between 25–35-year-olds and 50–60-year-olds, we examine each of the 49 object-comparison dimensions defined by the prior interpretable model. These dimensions are human-interpretable, quantifying similarities such as “metallic” or “food-related.” We modify the architecture of that model to learn the preferences of each age group for each of those dimensions. We then compare those preferences between each age group.

  For the second task, between six-year-olds and adults, we examine the classes of object-comparison relations described by taxonomic and thematic reasoning. We use the interpretable model to assign taxonomic and thematic scores to object-comparison questions, then select questions from amongst those to administer to six-year-olds. Finally, we contrast their responses with previous adult ones to elicit age-related changes in the prioritization of taxonomic and thematic reasoning, both in absolute terms and relative to one another.

  In the context of prior literature, we provide measures of differences in object-similarity judgment between younger and older adults for each of 49 fine-grained types of object similarity and remark upon the resulting trends. We corroborate a thematic-to-taxonomic trend in thinking from adolescence to adulthood. Finally, we expand the knowledge-base of the common-resource THINGS initiative with our results.

• Subjects / Keywords

  • age
  • object similarity
  • similarity
  • relation
  • object concept
  • embedding
  • computational
  • neural network
  • adult
  • child
  • children
  • NLP
  • natural language processing
  • taxonomic
  • thematic
  • THINGS initiative

• Graduation date

  Fall 2022

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/r3-mz1n-v607

• License

  This thesis is made available by the University of Alberta Library with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.