FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2163097513> ?p ?o ?g. }

• W2163097513 abstract "Ad-hoc Meaning Substitution Causes Shape and Material Biases: Computational Explanation for Emergence of Word Learning Biases Kosuke Kurosaki (kurosaki@complex.eng.hokudai.ac.jp) Graduate School of Information Science and Technology, Hokkaido University, Sapporo, Hokkaido, JAPAN Takashi Omori (omori@lab.tamagawa.ac.jp) Intelligent Robot Research Center, Tamagawa University Research Institute, Tokyo, JAPAN Abstract We illustrated a computational mechanism of shape bias and material bias with “ad-hoc meaning substitution (AMS)” hypothesis and verified it by computer simu- lations. AMS represents that when given a novel word and a instance object/substance, children substitute a known noun meaning nearest to the instance and the instance itself as an ad-hoc template of the novel noun meaning. The substitution enables fast mapping and subsequent vocabulary spurt. To describe internal pro- cess of AMS, we introduced “word distributional pro- totype (WDP)” as an explicit representation of word meaning with an inductive learning function. Simula- tion 1 revealed that when neural networks with WDP and AMS were given biased vocabularies reflecting those of young children, it demonstrated shape, material, and overgeneralized shape biases which means wrong shape bias over material bias. This result suggests that the triad of word meaning induction, AMS, and early biased vocabulary is essential for emergence of the biases. Sim- ulation 2 introduced a notion of maturity that denote a degree of learning convergence for each word meaning, and then networks showed neither shape nor material bias during early small vocabulary. This result indicates that the age of bias emerges is decided by the matu- rity. These results suggest that phenomena concerning shape and material biases are explicable with the simple ad-hoc learning mechanism instead of meta learning or built-in language-specific ones. (Dickinson, 1988; Soja et al., 1991). Those biases are investigated by novel noun generalization task (Samuel- son, 2002; Samuelson & Smith, 1999). In the task, ex- perimenter prepares solid and nonsolid sets. Each set consists of three stimuli: a novel target stimulus, shape- match stimuli that have the same shape as the target stimulus, and material-match stimuli that have the same material . First, an experimenter assigns a novel noun to the target stimulus in front of a subject. Next, the ex- perimenter presents a corresponding shape-match stim- ulus and a material-match stimulus together and asks the subject to select one that can be called by the same noun as the target stimulus. When either is selected sig- nificantly more often by some subjects, we conclude that subjects have a bias to generalize novel nouns based on similarity in shape or material. Shape bias is a behavior that when a novel solid tar- get stumulus is named with a novel noun, people tend to extend the noun to shape-match stimulus (Landau, Smith, & Jones, 1988). Material bias is a behavior that when a novel nonsolid target stimulus is named with a novel noun, they tend to generalize the noun to the material-match stimulus (Dickinson, 1988; Soja et al., 1991). We can summarize some experimental findings as below: (1.1.1a) shape bias doesn’t appear to solid stimuli when subjects have small vocabulary (Samuelson & Smith, 1999; Smith, 1995); and (1.1.1b) it does after middle vocabulary (Landau, Smith, & Jones, 1988; Soja et al., 1991; Dickinson, 1988; Imai & Gentner, 1997); (1.1.2a) material bias doesn’t appear during small vo- cabulary (Samuelson & Smith, 1999); or (1.1.2b) “over- generalized shape bias,” which means novel noun named to nonsolid target stimulus is extended also to shape- match stimuli, happens to appear during small vocabu- lary (Samuelson, 2002); and (1.1.2c) material bias ap- pears after sufficiently large vocabulary (Soja et al., 1991; Dickinson, 1988). From these findings, we argue that the biases result from simple learning (“learned bias account [LBA]”) be- cause: (1.2a) The fact that shape and material biases appear after children have acquired certain number of words indicates that they emerge as a consequence of vocabulary learning; (1.2b) emergence of the overgen- eralized shape bias suggests that both biases consist of common mechanism instead of separate modules and it causes the overgeneralization; (1.2c) both biases arise in almost identical situations except for the solidity of tar- Introduction When we encounter a novel word such as Gavagai and guess its meaning, too many logically possible meanings exist (Quine, 1960). Nevertheless, children can estimate words meanings very fast (Carey & Bartlett, 1978). Such fast mapping can’t be explained by existing machine learning algorithms based on trial and error. The ability appears after child’s productive noun vocabulary exceeds about 50 words, and the vocabulary starts to grow quite rapidly(vocabulary spurt). To explain these phenomena, developmental psychol- ogists have suggested word learning biases. When ap- plying a novel name to an object, the biases make chil- dren focus on particular features instead of other possible features and estimate words’ meaning accurately (e.g. Landau, Smith, & Jones, 1988; Soja, Carey, & Spelke, 1991). Problem is that these biases are just phenomeno- logical explanations that can’t explain why they exist or how they are processed in human brain. In this paper, we try to illustrate a computational mechanism of shape bias (Landau, Smith, & Jones, 1988) and material bias" @default.
• W2163097513 created "2016-06-24" @default.
• W2163097513 creator A5017743835 @default.
• W2163097513 creator A5049494554 @default.
• W2163097513 date "2006-01-01" @default.
• W2163097513 modified "2023-09-26" @default.
• W2163097513 title "Ad-hoc Meaning Substitution Causes Shape and Material Biases : Computational Explanation for Emergence of Word Learning Biases" @default.
• W2163097513 cites W1679913846 @default.
• W2163097513 cites W1980491396 @default.
• W2163097513 cites W1982374456 @default.
• W2163097513 cites W1986758797 @default.
• W2163097513 cites W1993819511 @default.
• W2163097513 cites W2000255081 @default.
• W2163097513 cites W2001850074 @default.
• W2163097513 cites W2035122010 @default.
• W2163097513 cites W2036529451 @default.
• W2163097513 cites W2037667353 @default.
• W2163097513 cites W2040461547 @default.
• W2163097513 cites W2078787448 @default.
• W2163097513 cites W2125500543 @default.
• W2163097513 cites W2144209400 @default.
• W2163097513 cites W2148283460 @default.
• W2163097513 cites W2028520710 @default.
• W2163097513 hasPublicationYear "2006" @default.
• W2163097513 type Work @default.
• W2163097513 sameAs 2163097513 @default.
• W2163097513 citedByCount "0" @default.
• W2163097513 crossrefType "journal-article" @default.
• W2163097513 hasAuthorship W2163097513A5017743835 @default.
• W2163097513 hasAuthorship W2163097513A5049494554 @default.
• W2163097513 hasConcept C121934690 @default.
• W2163097513 hasConcept C138885662 @default.
• W2163097513 hasConcept C154945302 @default.
• W2163097513 hasConcept C15744967 @default.
• W2163097513 hasConcept C199360897 @default.
• W2163097513 hasConcept C204321447 @default.
• W2163097513 hasConcept C2777601683 @default.
• W2163097513 hasConcept C2778220771 @default.
• W2163097513 hasConcept C2780876879 @default.
• W2163097513 hasConcept C41008148 @default.
• W2163097513 hasConcept C41895202 @default.
• W2163097513 hasConcept C542102704 @default.
• W2163097513 hasConcept C90805587 @default.
• W2163097513 hasConceptScore W2163097513C121934690 @default.
• W2163097513 hasConceptScore W2163097513C138885662 @default.
• W2163097513 hasConceptScore W2163097513C154945302 @default.
• W2163097513 hasConceptScore W2163097513C15744967 @default.
• W2163097513 hasConceptScore W2163097513C199360897 @default.
• W2163097513 hasConceptScore W2163097513C204321447 @default.
• W2163097513 hasConceptScore W2163097513C2777601683 @default.
• W2163097513 hasConceptScore W2163097513C2778220771 @default.
• W2163097513 hasConceptScore W2163097513C2780876879 @default.
• W2163097513 hasConceptScore W2163097513C41008148 @default.
• W2163097513 hasConceptScore W2163097513C41895202 @default.
• W2163097513 hasConceptScore W2163097513C542102704 @default.
• W2163097513 hasConceptScore W2163097513C90805587 @default.
• W2163097513 hasIssue "28" @default.
• W2163097513 hasLocation W21630975131 @default.
• W2163097513 hasOpenAccess W2163097513 @default.
• W2163097513 hasPrimaryLocation W21630975131 @default.
• W2163097513 hasRelatedWork W1705431365 @default.
• W2163097513 hasRelatedWork W1970587543 @default.
• W2163097513 hasRelatedWork W2004108411 @default.
• W2163097513 hasRelatedWork W2006170604 @default.
• W2163097513 hasRelatedWork W2043632556 @default.
• W2163097513 hasRelatedWork W2088397639 @default.
• W2163097513 hasRelatedWork W2144713825 @default.
• W2163097513 hasRelatedWork W2152367440 @default.
• W2163097513 hasRelatedWork W2199101053 @default.
• W2163097513 hasRelatedWork W2222340070 @default.
• W2163097513 hasRelatedWork W2229021201 @default.
• W2163097513 hasRelatedWork W2400674239 @default.
• W2163097513 hasRelatedWork W2513866468 @default.
• W2163097513 hasRelatedWork W2766891600 @default.
• W2163097513 hasRelatedWork W2767390558 @default.
• W2163097513 hasRelatedWork W2767703312 @default.
• W2163097513 hasRelatedWork W3034509676 @default.
• W2163097513 hasRelatedWork W3100361495 @default.
• W2163097513 hasRelatedWork W3111485364 @default.
• W2163097513 hasRelatedWork W161659516 @default.
• W2163097513 hasVolume "28" @default.
• W2163097513 isParatext "false" @default.
• W2163097513 isRetracted "false" @default.
• W2163097513 magId "2163097513" @default.
• W2163097513 workType "article" @default.