FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W34606440> ?p ?o ?g. }

• W34606440 abstract "At the heart of nearly every model in cognitive science and arti cial intelligence is the notion of roles and their llers. Roles appear as explicit slots in frame systems (Minsky, 1975), as particular positions in predicate calculus representations, as role-speci c groups in some connectionist models (McClelland & Kawamoto, 1986), and as distributed patterns in role space in other connectionist models (Smolensky, 1990). But where do roles come from? In most systems, this is not treated as an issue; objects are simply assigned fully formed structured representations. But for some problems the decomposition of an object into a set of roles and the assigning of llers to these roles appears to be half of the battle. For instance, systems attempting to make analogies between two domains are already half nished before they start if they are given sets of roles and associated llers for the items being compared (Holyoak & Thagard, 1989). We believe that problems such as analogy can be addressed adequately only when the structure of representations is grounded in perception and action (Harnad, 1991). Recent e orts to ground semantics in visual input (Bartell & Cottrell, 1991; Regier, 1990) have focused on what distinguishes predicates such as fast from slow or above from below. These approaches have not looked at questions of how objects are assigned to di erent roles or how a system would cope with predicates taking di erent numbers of arguments. Consider a system which is given a visual display containing a number of objects and the task of describing what it sees. It must do several things. It must nd one or more objects, assign labels of some sort to them, recognize that a relation exists between two or more of the objects (if it does), assign a label to the relation, and assign the objects to the roles of the relation. The assignment of objects to roles depends on the relative salience of the objects; in each relation the most salient object becomes the trajector (Langacker, 1987), realized linguistically as the subject; the other objects are landmarks, realized as the direct object or some other argument. This view recognizes the non-equivalence of pairs such as the square is above the circle and the circle is below the square; they di er with respect to the perceived relative salience of the two objects. One possibility for how this might be done is as follows. The system scans the visual input for objects, identifying them as it locates them. It keeps track of the sequence of scanning actions it makes, and when two or more objects are perceived in succession, the memory for the recent" @default.
• W34606440 created "2016-06-24" @default.
• W34606440 creator A5056681764 @default.
• W34606440 creator A5069886855 @default.
• W34606440 date "1992-01-01" @default.
• W34606440 modified "2023-09-27" @default.
• W34606440 title "Grounding via Scanning: Cooking up Roles from Scratch" @default.
• W34606440 cites W2013494846 @default.
• W34606440 cites W2086461330 @default.
• W34606440 cites W2105048688 @default.
• W34606440 cites W2110485445 @default.
• W34606440 cites W2156159857 @default.
• W34606440 cites W2175268987 @default.
• W34606440 hasPublicationYear "1992" @default.
• W34606440 type Work @default.
• W34606440 sameAs 34606440 @default.
• W34606440 citedByCount "2" @default.
• W34606440 crossrefType "journal-article" @default.
• W34606440 hasAuthorship W34606440A5056681764 @default.
• W34606440 hasAuthorship W34606440A5069886855 @default.
• W34606440 hasConcept C111472728 @default.
• W34606440 hasConcept C12713177 @default.
• W34606440 hasConcept C138885662 @default.
• W34606440 hasConcept C140146324 @default.
• W34606440 hasConcept C154945302 @default.
• W34606440 hasConcept C15744967 @default.
• W34606440 hasConcept C169760540 @default.
• W34606440 hasConcept C177264268 @default.
• W34606440 hasConcept C188147891 @default.
• W34606440 hasConcept C199360897 @default.
• W34606440 hasConcept C26760741 @default.
• W34606440 hasConcept C2781238097 @default.
• W34606440 hasConcept C41008148 @default.
• W34606440 hasConcept C50644808 @default.
• W34606440 hasConcept C521332185 @default.
• W34606440 hasConcept C8521452 @default.
• W34606440 hasConceptScore W34606440C111472728 @default.
• W34606440 hasConceptScore W34606440C12713177 @default.
• W34606440 hasConceptScore W34606440C138885662 @default.
• W34606440 hasConceptScore W34606440C140146324 @default.
• W34606440 hasConceptScore W34606440C154945302 @default.
• W34606440 hasConceptScore W34606440C15744967 @default.
• W34606440 hasConceptScore W34606440C169760540 @default.
• W34606440 hasConceptScore W34606440C177264268 @default.
• W34606440 hasConceptScore W34606440C188147891 @default.
• W34606440 hasConceptScore W34606440C199360897 @default.
• W34606440 hasConceptScore W34606440C26760741 @default.
• W34606440 hasConceptScore W34606440C2781238097 @default.
• W34606440 hasConceptScore W34606440C41008148 @default.
• W34606440 hasConceptScore W34606440C50644808 @default.
• W34606440 hasConceptScore W34606440C521332185 @default.
• W34606440 hasConceptScore W34606440C8521452 @default.
• W34606440 hasLocation W346064401 @default.
• W34606440 hasOpenAccess W34606440 @default.
• W34606440 hasPrimaryLocation W346064401 @default.
• W34606440 hasRelatedWork W109870594 @default.
• W34606440 hasRelatedWork W132195071 @default.
• W34606440 hasRelatedWork W2019975382 @default.
• W34606440 hasRelatedWork W2135719810 @default.
• W34606440 hasRelatedWork W2476995767 @default.
• W34606440 hasRelatedWork W2569483224 @default.
• W34606440 hasRelatedWork W2984434901 @default.
• W34606440 hasRelatedWork W308551192 @default.
• W34606440 hasRelatedWork W60533568 @default.
• W34606440 isParatext "false" @default.
• W34606440 isRetracted "false" @default.
• W34606440 magId "34606440" @default.
• W34606440 workType "article" @default.