Bayesian model ensembling using meta-trained recurrent neural networks

There is no fulltext present in this item.

Title:	Bayesian model ensembling using meta-trained recurrent neural networks
Author(s):	Ambrogioni, L. ; Berezutskaya, Y. ; Güçlü, U. ; Borne, E.W.P. van den; Güçlütürk, Y. ; Gerven, M.A.J. van ; Maris, E.G.G.
Publication year:	2017
In:	Workshop on Meta-Learning (MetaLearn 2017) - Neural Information Processing Systems (NIPS 2017), pp. 1-5
Publisher:	[S.l. : s.n.]
Related links:	http://metalearning.ml/2017/papers/metalearn17_ambrogioni.pdf
Annotation:	31st Conference on Neural Information Processing Systems (NIPS 2017) (Long Beach, CA, USA, December 4-9, 2017)
Publication type:	Article in monograph or in proceedings
Please use this identifier to cite or link to this item : https://hdl.handle.net/2066/180405
Display more details
Subject:	Action, intention, and motor control Cognitive artificial intelligence DI-BCB_DCC_Theme 2: Perception, Action and Control DI-BCB_DCC_Theme 4: Brain Networks and Neuronal Communication
Organization:	SW OZ DCC SMN SW OZ DCC KI SW OZ DCC CO
Book title:	Workshop on Meta-Learning (MetaLearn 2017) - Neural Information Processing Systems (NIPS 2017)
Page start:	p. 1
Page end:	p. 5
Abstract:	In this paper we demonstrate that a recurrent neural network meta-trained on an ensemble of arbitrary classification tasks can be used as an approximation of the Bayes optimal classifier. This result is obtained by relying on the framework of e-free approximate Bayesian inference, where the Bayesian posterior is approximated by training a neural network using synthetic samples. We denote the resulting model as neural ensembler. We show that a single neural ensembler trained on a large set of synthetic data achieves competitive classification performance on multiple real-world classification problems without additional training.
Languages used:	English (eng)