A Competitive Modular Connectionist Architecture

By Robert A. Jacobs, Michael I. Jordan. 1990

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Towards causal generative scene models via competition of experts (2020, methodology): “To achieve specialisation on different object classes in our model, we build on ideas from previous work using competitive training of experts (Jacobs & Jordan, 1991).”
Knowledge Transfer in Deep Block-Modular Neural Networks (2015, background): “al tasks has been addressed in modular neural networks. These networks are able to detect dierent training patterns, corresponding to dierent tasks, and allocate dierent sub-networks to learn them [15,19]. In the current paper we explore an alternative approach to training DNNs, partly inspired by m”
Transfer of learning by composing solutions of elemental sequential tasks (2004, background): “The best interpretation of this architecture is as an \associative Gaussian mixture model" (see Jacobs & Jordan, 1991 ; Jacobs et al. 1991 ) of the distribution of training patterns.”
Modular Neural Networks and Self-Decomposition (1997, methodology): “Jacobs, R. and M.I. Jordan (1993).”
Bias/Variance Analyses of Mixtures-of-Experts Architectures (1997, background): “(1991), Jacobs and Jordan (1991, 1993), Jordan and Jacobs ( Peng et al .”
Evolutionary ensembles with negative correlation learning (2000, methodology): “Since the early 1990's, algorithms based on similar ideas have been developed in many di erent but related forms, such as NN ensembles [5, 2], mixtures of experts [6, 7, 8, 9], and various boosting and bagging methods [10, 11, 12].”