scholarly journals Bayesian Spatial Inversion and Conjugate Selection Gaussian Prior Models

2021 ◽  
Vol 9 (2) ◽  
pp. 420-445
Author(s):  
Henning Omre ◽  
Kjartan Rimstad
Keyword(s):  
Gaussian Prior ◽  
Prior Models

scholarly journals Fast Low-Rank Bayesian Matrix Completion With Hierarchical Gaussian Prior Models

2018 ◽  
Vol 66 (11) ◽  
pp. 2804-2817 ◽  
Author(s):  
Linxiao Yang ◽  
Jun Fang ◽  
Huiping Duan ◽  
Hongbin Li ◽  
Bing Zeng
Keyword(s):  
Matrix Completion ◽  
Low Rank ◽  
Gaussian Prior ◽  
Prior Models

The Embodiment of Concepts

2018 ◽  
pp. 640-660
Author(s):  
Michiel Van Elk ◽  
Harold Bekkering
Keyword(s):  
Prediction Error ◽  
Predictive Processing ◽  
Prediction Errors ◽  
Conceptual Representation ◽  
Recurrent Processing ◽  
Different Dimensions ◽  
Prior Models ◽  
Processing Framework

We characterize theories of conceptual representation as embodied, disembodied, or hybrid according to their stance on a number of different dimensions: the nature of concepts, the relation between language and concepts, the function of concepts, the acquisition of concepts, the representation of concepts, and the role of context. We propose to extend an embodied view of concepts, by taking into account the importance of multimodal associations and predictive processing. We argue that concepts are dynamically acquired and updated, based on recurrent processing of prediction error signals in a hierarchically structured network. Concepts are thus used as prior models to generate multimodal expectations, thereby reducing surprise and enabling greater precision in the perception of exemplars. This view places embodied theories of concepts in a novel predictive processing framework, by highlighting the importance of concepts for prediction, learning and shaping categories on the basis of prediction errors.

Sign in / Sign up

Export Citation Format

Share Document