FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W52677385> ?p ?o ?g. }

• W52677385 abstract "Modeling sequential information integration with parallel constraint satisfaction Katja Mehlhorn (katja.mehlhorn@phil.tu-chemnitz.de) Chemnitz University of Technology, Department of Psychology Georg Jahn (georg.jahn@uni-greifswald.de) University of Greifswald, Department of Psychology Abstract An important aspect of human cognition is the sequential integration of observations while striving for a consistent mental representation. Recent research consistently stresses the importance of fast automatic processes for integrating information available at a certain point in time. However, it is not clear, how such processes allow for maintaining a consistent and up to date mental representation in the light of new information. We compare variants of two methods of modeling sequential information integration with parallel constraint satisfaction models: carrying over results from the previous integration step or decaying input strength of older observations. Results of these models for consistent and inconsistent sets of observations are compared to human data from a diagnostic reasoning task. Keywords: information integration, belief diagnostic reasoning, memory activation, satisfaction modeling updating, constraint Introduction A key feature of many everyday reasoning tasks is that observations are processed sequentially. Whether it is in diagnostic reasoning, in decision making, or in belief updating, often information becomes available step by step. If a large amount of information is given all at once, it might only be perceived and understood sequentially due to limited cognitive capacities. Although possible implications of the sequential nature of tasks (e.g., order effects) have been discussed (e.g., Hogarth & Einhorn, 1992; Wang, Johnson, & Zhang, 2006), the underlying cognitive mechanisms are not fully understood. Recent research consistently points out the importance of fast automatic processes for integrating information available at a certain point in time (e.g. Glockner & Betsch, 2008). However, it is not clear how such processes allow for maintaining a consistent mental representation in the light of new incoming information. In this paper, we explore alternative implementations of such processes in connectionist constraint satisfaction models. Previous research has shown that reasoners hold knowledge structures that reflect the structure of the task in the environment (e.g., Anderson, 1983; Gigerenzer, Hoffrage, & Kleinbolting, 1991; Thomas, Dougherty, Sprenger, & Harbison, 2008). For example, a physician learns, with an increasing number of patients encountered, which symptoms are associated with which diseases and how strong these associations are. Given such an adapted knowledge structure, observations can serve as a cue for the retrieval of associated knowledge from long-term memory (e.g., Kintsch, 1998; Thomas et al., 2008; Baumann, Mehlhorn, & Bocklisch, 2007). To maintain a consistent representation of the task at hand, this newly activated information somehow needs to be integrated with previous observations and previously activated knowledge. How is this achieved? Wang et al. (2006) have proposed a connectionist model of sequential integration based on the idea of explanatory coherence that, probably most prominently, was introduced by Thagard (1989, 2000) in the field of scientific discovery. Thagard implemented explanatory coherence among interconnected propositions in a connectionist constraint satisfaction model (ECHO). In ECHO, propositions are represented by nodes. The nodes are interconnected by symmetric excitatory and inhibitory links representing the relations (constraints) between them. Nodes representing observed information are additionally connected to a special activation node (special evidence unit = SEU), which always has an activation value of 1 and is the model’s “energy source”. Connecting not all, but only these data nodes to the energy source reflects the idea that empirical data are weighted more strongly than theoretical hypotheses held by the reasoner (Thagard, 1989). The strength of a proposition in the network is indicated by the numerical activation of its node. Before the network is integrated, activation of all nodes is set to default values. Then, activation spreads from the SEU to the data nodes and then to other connected nodes. The net input each node receives is calculated as the weighted sum of the activation of all nodes it is connected to. After calculating the input for each node, the activation of all nodes is updated synchronously. These two steps are repeated iteratively, until activation stops changing substantially. The more consistent a proposition is with the observed information and other related propositions, the higher is the activation of its node when the network settles. The idea of constraint satisfaction has been widely applied to areas such as text comprehension (Kintsch, 1998), social impression formation (Thagard & Kunda, 1998), visuo-spatial reasoning (Thagard & Shelley, 1997), causal reasoning (Hagmeyer & Waldmann, 2002), medical diagnosis (Arocha & Patel, 1995), and decision making (Glockner & Betsch, 2008). In all of these different tasks, reasoners need to find an interpretation that is more coherent with the available information than possible alternative interpretations. Such coherent interpretations can be the meaning of a word that fits best in the current context, the impression about a person that is most coherent with one’s previous impression about him/her, or it can be the diagnosis that best explains the set of a patient’s symptoms. Applied successfully to model various phenomena in all the above domains, constraint satisfaction models have been described as a “computationally efficient approximation to" @default.
• W52677385 created "2016-06-24" @default.
• W52677385 creator A5018004853 @default.
• W52677385 creator A5039462142 @default.
• W52677385 date "2009-01-01" @default.
• W52677385 modified "2023-09-23" @default.
• W52677385 title "Modeling sequential information integration with parallel constraint satisfaction" @default.
• W52677385 cites W1674947250 @default.
• W52677385 cites W1696057967 @default.
• W52677385 cites W169684159 @default.
• W52677385 cites W177597204 @default.
• W52677385 cites W1803467130 @default.
• W52677385 cites W1979410064 @default.
• W52677385 cites W1990948551 @default.
• W52677385 cites W2048175187 @default.
• W52677385 cites W2082557450 @default.
• W52677385 cites W2084809649 @default.
• W52677385 cites W2150234857 @default.
• W52677385 cites W2153162913 @default.
• W52677385 cites W2767524492 @default.
• W52677385 cites W2913008767 @default.
• W52677385 cites W3125306911 @default.
• W52677385 cites W594310934 @default.
• W52677385 cites W83607462 @default.
• W52677385 cites W94599463 @default.
• W52677385 hasPublicationYear "2009" @default.
• W52677385 type Work @default.
• W52677385 sameAs 52677385 @default.
• W52677385 citedByCount "4" @default.
• W52677385 countsByYear W526773852012 @default.
• W52677385 countsByYear W526773852015 @default.
• W52677385 countsByYear W526773852020 @default.
• W52677385 crossrefType "journal-article" @default.
• W52677385 hasAuthorship W52677385A5018004853 @default.
• W52677385 hasAuthorship W52677385A5039462142 @default.
• W52677385 hasConcept C119857082 @default.
• W52677385 hasConcept C124101348 @default.
• W52677385 hasConcept C127413603 @default.
• W52677385 hasConcept C138885662 @default.
• W52677385 hasConcept C154945302 @default.
• W52677385 hasConcept C15744967 @default.
• W52677385 hasConcept C161407221 @default.
• W52677385 hasConcept C169760540 @default.
• W52677385 hasConcept C169900460 @default.
• W52677385 hasConcept C17744445 @default.
• W52677385 hasConcept C180747234 @default.
• W52677385 hasConcept C188147891 @default.
• W52677385 hasConcept C199539241 @default.
• W52677385 hasConcept C199622910 @default.
• W52677385 hasConcept C201995342 @default.
• W52677385 hasConcept C2524010 @default.
• W52677385 hasConcept C2776036281 @default.
• W52677385 hasConcept C2776359362 @default.
• W52677385 hasConcept C2776401178 @default.
• W52677385 hasConcept C2780451532 @default.
• W52677385 hasConcept C33326189 @default.
• W52677385 hasConcept C33923547 @default.
• W52677385 hasConcept C41008148 @default.
• W52677385 hasConcept C41895202 @default.
• W52677385 hasConcept C44616089 @default.
• W52677385 hasConcept C49937458 @default.
• W52677385 hasConcept C94625758 @default.
• W52677385 hasConceptScore W52677385C119857082 @default.
• W52677385 hasConceptScore W52677385C124101348 @default.
• W52677385 hasConceptScore W52677385C127413603 @default.
• W52677385 hasConceptScore W52677385C138885662 @default.
• W52677385 hasConceptScore W52677385C154945302 @default.
• W52677385 hasConceptScore W52677385C15744967 @default.
• W52677385 hasConceptScore W52677385C161407221 @default.
• W52677385 hasConceptScore W52677385C169760540 @default.
• W52677385 hasConceptScore W52677385C169900460 @default.
• W52677385 hasConceptScore W52677385C17744445 @default.
• W52677385 hasConceptScore W52677385C180747234 @default.
• W52677385 hasConceptScore W52677385C188147891 @default.
• W52677385 hasConceptScore W52677385C199539241 @default.
• W52677385 hasConceptScore W52677385C199622910 @default.
• W52677385 hasConceptScore W52677385C201995342 @default.
• W52677385 hasConceptScore W52677385C2524010 @default.
• W52677385 hasConceptScore W52677385C2776036281 @default.
• W52677385 hasConceptScore W52677385C2776359362 @default.
• W52677385 hasConceptScore W52677385C2776401178 @default.
• W52677385 hasConceptScore W52677385C2780451532 @default.
• W52677385 hasConceptScore W52677385C33326189 @default.
• W52677385 hasConceptScore W52677385C33923547 @default.
• W52677385 hasConceptScore W52677385C41008148 @default.
• W52677385 hasConceptScore W52677385C41895202 @default.
• W52677385 hasConceptScore W52677385C44616089 @default.
• W52677385 hasConceptScore W52677385C49937458 @default.
• W52677385 hasConceptScore W52677385C94625758 @default.
• W52677385 hasIssue "31" @default.
• W52677385 hasLocation W526773851 @default.
• W52677385 hasOpenAccess W52677385 @default.
• W52677385 hasPrimaryLocation W526773851 @default.
• W52677385 hasRelatedWork W1520942242 @default.
• W52677385 hasRelatedWork W1576457731 @default.
• W52677385 hasRelatedWork W1594662628 @default.
• W52677385 hasRelatedWork W1778002585 @default.
• W52677385 hasRelatedWork W1976692131 @default.
• W52677385 hasRelatedWork W1979410064 @default.
• W52677385 hasRelatedWork W2038901583 @default.

• next