FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W210231100> ?p ?o ?g. }

• W210231100 abstract "The Role of Familiarity, Priming and Perception in Similarity Judgments Laura M. Hiatt and J. Gregory Trafton U.S. Naval Research Laboratory Washington, DC 20375 USA {laura.hiatt, greg.trafton}@nrl.navy.mil Abstract no clear prototype). The experiment showed a striking asym- metry in similarity judgments between the different colors when they were presented with different frequencies during an irrelevant training task: colors which had been trained on less often were considered more similar to colors which had been trained on more often than the other way around. To account for these low-level results, Polk et al. (2002) imple- mented a neural network which simulated the asymmetry by measuring the ease with which the network switches between different activation patterns; those that are more stable (e.g., high-frequency patterns) were easier to assimilate to. In our approach, we match the human subject data from Polk et al. (2002) while attempting to address three additional points. First, there was a second significant effect, that the ratings in general increased over time, that the above models do not address. Second, we wanted our approach to provide explicit cognitive processes for similarity ratings. Third, we believe that inherent perceptual similarity also plays a part in these types of similarity judgments (e.g., purple is inherently more similar to blue than to orange) (Smith & Heise, 1992). To this end, we begin our approach with the cognitive ar- chitecture ACT-R (Adaptive Characterization of Thought – Rational) (Anderson, 2007). Using ACT-R, we account for similarity judgments by considering three values provided a priori by the architecture. The first, familiarity, is represented as a base-level activation value of a concept, which represents its frequency and recency of use. The second, priming, is based on spreading activation, which disperses activation be- tween different, associated concepts in declarative memory (Anderson, 1983; Harrison & Trafton, 2010). In addition, we utilize an extension to ACT-R which provides it with a cal- culation for measuring color similarity (Breslow, Ratwani, & Trafton, 2009; Breslow, Trafton, & Ratwani, 2009). Our model starts without any pre-existing declarative knowledge or network structures; all knowledge is created during the experiment. Over time, our cognitive model builds a network of concepts (e.g., color blocks) by learning associ- ations between them in the form of subsymbolic connections between their representations in declarative knowledge. Dur- ing each similarity judgment, the model combines its mea- sure of perceptual similarity with base-level and spreading activation to determine its response. On the first trial of an experiment, there is no spreading activation since there is no declarative knowledge and so the judgment is based purely on base-level activation and perceptual similarity; however, over time the model builds up declarative memories that may contribute to spreading activation in later trials. This explains the two main effects found in Polk et al.’s experiment. Dur- We present a novel way of accounting for similarity judgments. Our approach posits that similarity ratings stem from three main sources: familiarity, priming, and inherent perceptual similarity. We present a process model of our approach in the cognitive architecture ACT-R, and match our model’s predic- tions to data collected from a human subject experiment which involved simple perceptual stimuli. Familiarity accounts for rising ratings over time; priming accounts for asymmetric ef- fects that arise when the stimuli are shown with different fre- quencies. Pure perceptual similarity also predicts trends in the results. Overall, our model matched the data with R 2 of 0.99. Introduction Similarity is a critical and pervasive part of human cogni- tion (Medin, Goldstone, & Gentner, 1993). Similarity mea- sures are integral to object categorization and classification (Nosofsky, 1992). Similarity is also pervasive in problem solving (Novick, 1990), decision-making (Medin, Goldstone, & Markman, 1995), and memory (Roediger, 1990). As with many aspects of human cognition, however, the mechanisms that determine similarity are not yet fully understood. Var- ious theories abound, with none yet able to capture enough different types of situations to be called the winner (Rorissa, One interesting result in this field is asymmetries that have been shown to arise when making similarity judgments, even of very simple perceptual stimuli (Tversky, 1977; Rosch, 1975). Rosch (1975) argued that such similarity is based on mapping stimuli onto one another and, intuitively, non- prototypical stimuli map more easily onto prototypical stim- uli than vice versa. Tversky (1977) argued that is due to weighted feature matching, where the salience of features in the current context determines their weight; others agree with this thought in general (Medin et al., 1993; Glucksberg & Keysar, 1990). These two explanations, however, assume that either there is a clear prototype inherently present in the experiment (such as the more perceptually complex stimulus), or that stim- uli have various features which have a clear inherent or- der of cognitive preference and saliency (such as symmetry). They do not, however, provide any concrete explanations for why complexity or symmetry may lead to prototypicality or saliency. Polk, Behensky, Gonzalez, and Smith (2002) shed light on the situation by presenting an experiment that avoids the question entirely by using perceptual stimuli where the only feature was color (so there are no features to comparatively weigh), and where the color hues are fairly similar (so there is" @default.
• W210231100 created "2016-06-24" @default.
• W210231100 creator A5020141134 @default.
• W210231100 creator A5056412737 @default.
• W210231100 date "2013-01-01" @default.
• W210231100 modified "2023-09-23" @default.
• W210231100 title "The Role of Familiarity, Priming and Perception in Similarity Judgments" @default.
• W210231100 cites W1492367877 @default.
• W210231100 cites W1564157903 @default.
• W210231100 cites W1928882148 @default.
• W210231100 cites W1978976775 @default.
• W210231100 cites W1979104110 @default.
• W210231100 cites W2005081946 @default.
• W210231100 cites W2017483902 @default.
• W210231100 cites W2050375379 @default.
• W210231100 cites W2059975159 @default.
• W210231100 cites W2082557450 @default.
• W210231100 cites W2085904853 @default.
• W210231100 cites W2102162397 @default.
• W210231100 cites W2110071186 @default.
• W210231100 cites W2117259269 @default.
• W210231100 cites W2129922198 @default.
• W210231100 cites W2136966607 @default.
• W210231100 cites W2143364917 @default.
• W210231100 cites W2162566219 @default.
• W210231100 cites W2165934783 @default.
• W210231100 cites W2167723047 @default.
• W210231100 cites W2415747639 @default.
• W210231100 cites W2623695761 @default.
• W210231100 cites W2799095127 @default.
• W210231100 cites W1601333326 @default.
• W210231100 hasPublicationYear "2013" @default.
• W210231100 type Work @default.
• W210231100 sameAs 210231100 @default.
• W210231100 citedByCount "3" @default.
• W210231100 countsByYear W2102311002015 @default.
• W210231100 crossrefType "journal-article" @default.
• W210231100 hasAuthorship W210231100A5020141134 @default.
• W210231100 hasAuthorship W210231100A5056412737 @default.
• W210231100 hasConcept C100701293 @default.
• W210231100 hasConcept C103278499 @default.
• W210231100 hasConcept C115961682 @default.
• W210231100 hasConcept C154945302 @default.
• W210231100 hasConcept C15744967 @default.
• W210231100 hasConcept C169760540 @default.
• W210231100 hasConcept C169900460 @default.
• W210231100 hasConcept C180747234 @default.
• W210231100 hasConcept C26760741 @default.
• W210231100 hasConcept C41008148 @default.
• W210231100 hasConcept C59822182 @default.
• W210231100 hasConcept C81444415 @default.
• W210231100 hasConcept C86803240 @default.
• W210231100 hasConceptScore W210231100C100701293 @default.
• W210231100 hasConceptScore W210231100C103278499 @default.
• W210231100 hasConceptScore W210231100C115961682 @default.
• W210231100 hasConceptScore W210231100C154945302 @default.
• W210231100 hasConceptScore W210231100C15744967 @default.
• W210231100 hasConceptScore W210231100C169760540 @default.
• W210231100 hasConceptScore W210231100C169900460 @default.
• W210231100 hasConceptScore W210231100C180747234 @default.
• W210231100 hasConceptScore W210231100C26760741 @default.
• W210231100 hasConceptScore W210231100C41008148 @default.
• W210231100 hasConceptScore W210231100C59822182 @default.
• W210231100 hasConceptScore W210231100C81444415 @default.
• W210231100 hasConceptScore W210231100C86803240 @default.
• W210231100 hasIssue "35" @default.
• W210231100 hasLocation W2102311001 @default.
• W210231100 hasOpenAccess W210231100 @default.
• W210231100 hasPrimaryLocation W2102311001 @default.
• W210231100 hasRelatedWork W1928882148 @default.
• W210231100 hasRelatedWork W2001548857 @default.
• W210231100 hasRelatedWork W2056336939 @default.
• W210231100 hasRelatedWork W2070984693 @default.
• W210231100 hasRelatedWork W2082557450 @default.
• W210231100 hasRelatedWork W2083064751 @default.
• W210231100 hasRelatedWork W2085293874 @default.
• W210231100 hasRelatedWork W2099550752 @default.
• W210231100 hasRelatedWork W2102162397 @default.
• W210231100 hasRelatedWork W2110071186 @default.
• W210231100 hasRelatedWork W2128112785 @default.
• W210231100 hasRelatedWork W2147539805 @default.
• W210231100 hasRelatedWork W2165934783 @default.
• W210231100 hasRelatedWork W2327810546 @default.
• W210231100 hasRelatedWork W2485764162 @default.
• W210231100 hasRelatedWork W2766069215 @default.
• W210231100 hasRelatedWork W2766430251 @default.
• W210231100 hasRelatedWork W2768067805 @default.
• W210231100 hasRelatedWork W3140498427 @default.
• W210231100 hasRelatedWork W3185925200 @default.
• W210231100 hasVolume "35" @default.
• W210231100 isParatext "false" @default.
• W210231100 isRetracted "false" @default.
• W210231100 magId "210231100" @default.
• W210231100 workType "article" @default.