Cover Image

A deep learning method for comparing Bayesian hierarchical models

Bayesian model comparison (BMC) offers a principled approach to assessing the relative merits of competing computational models and propagating uncertainty into model selection decisions. However, BMC is often intractable for the popular class of hierarchical models due to their high-dimensional nes...

Full description

Saved in:
Bibliographic Details
Main Authors: Elsemüller, Lasse (Author) , Schnürch, Martin (Author) , Bürkner, Paul-Christian (Author) , Radev, Stefan (Author)
Format: Article (Journal)
Language:English
Published: 2024
In: Psychological methods
Year: 2024, Pages: ?
ISSN:1939-1463
DOI:10.1037/met0000645
Online Access:Verlag, lizenzpflichtig, Volltext: https://doi.org/10.1037/met0000645
Get full text
Author Notes:Lasse Elsemüller, Martin Schnuerch, Paul-Christian Bürkner, Stefan T. Radev

MARC

LEADER 00000caa a22000002c 4500
001 1903735750
003 DE-627
005 20241205174752.0
007 cr uuu---uuuuu
008 240930s2024 xx |||||o 00| ||eng c
024 7 |a 10.1037/met0000645  |2 doi 
035 |a (DE-627)1903735750 
035 |a (DE-599)KXP1903735750 
035 |a (OCoLC)1475313111 
040 |a DE-627  |b ger  |c DE-627  |e rda 
041 |a eng 
084 |a 11  |2 sdnb 
100 1 |a Elsemüller, Lasse  |e VerfasserIn  |0 (DE-588)1166863891  |0 (DE-627)1030747598  |0 (DE-576)510926959  |4 aut 
245 1 2 |a A deep learning method for comparing Bayesian hierarchical models  |c Lasse Elsemüller, Martin Schnuerch, Paul-Christian Bürkner, Stefan T. Radev 
264 1 |c 2024 
300 |a ? 
336 |a Text  |b txt  |2 rdacontent 
337 |a Computermedien  |b c  |2 rdamedia 
338 |a Online-Ressource  |b cr  |2 rdacarrier 
500 |a Gesehen am 30.09.2024 
520 |a Bayesian model comparison (BMC) offers a principled approach to assessing the relative merits of competing computational models and propagating uncertainty into model selection decisions. However, BMC is often intractable for the popular class of hierarchical models due to their high-dimensional nested parameter structure. To address this intractability, we propose a deep learning method for performing BMC on any set of hierarchical models which can be instantiated as probabilistic programs. Since our method enables amortized inference, it allows efficient re-estimation of posterior model probabilities and fast performance validation prior to any real-data application. In a series of extensive validation studies, we benchmark the performance of our method against the state-of-the-art bridge sampling method and demonstrate excellent amortized inference across all BMC settings. We then showcase our method by comparing four hierarchical evidence accumulation models that have previously been deemed intractable for BMC due to partly implicit likelihoods. Additionally, we demonstrate how transfer learning can be leveraged to enhance training efficiency. We provide reproducible code for all analyses and an open-source implementation of our method. (PsycInfo Database Record (c) 2024 APA, all rights reserved) 
650 4 |a Bayesian Analysis 
650 4 |a Deep Neural Networks 
650 4 |a Inference 
650 4 |a Mathematical Modeling 
650 4 |a Probability 
650 4 |a Simulation 
700 1 |a Schnürch, Martin  |d 1991-  |e VerfasserIn  |0 (DE-588)1207475564  |0 (DE-627)1693803720  |4 aut 
700 1 |a Bürkner, Paul-Christian  |d 1991-  |e VerfasserIn  |0 (DE-588)1071571338  |0 (DE-627)826045715  |0 (DE-576)433154705  |4 aut 
700 1 |a Radev, Stefan  |d 1993-  |e VerfasserIn  |0 (DE-588)1155312392  |0 (DE-627)1016724993  |0 (DE-576)501536248  |4 aut 
773 0 8 |i Enthalten in  |t Psychological methods  |d Washington, DC : American Psychological Association, 1996  |g (2024), Seite ?  |h Online-Ressource  |w (DE-627)363738002  |w (DE-600)2103345-6  |w (DE-576)251938735  |x 1939-1463  |7 nnas  |a A deep learning method for comparing Bayesian hierarchical models 
773 1 8 |g year:2024  |g pages:?  |g extent:?  |a A deep learning method for comparing Bayesian hierarchical models 
856 4 0 |u https://doi.org/10.1037/met0000645  |x Verlag  |x Resolving-System  |z lizenzpflichtig  |3 Volltext 
951 |a AR 
992 |a 20240930 
993 |a Article 
994 |a 2024 
998 |g 1155312392  |a Radev, Stefan  |m 1155312392:Radev, Stefan  |p 4  |y j 
998 |g 1207475564  |a Schnürch, Martin  |m 1207475564:Schnürch, Martin  |p 2 
998 |g 1166863891  |a Elsemüller, Lasse  |m 1166863891:Elsemüller, Lasse  |d 100000  |d 100200  |e 100000PE1166863891  |e 100200PE1166863891  |k 0/100000/  |k 1/100000/100200/  |p 1  |x j 
999 |a KXP-PPN1903735750  |e 4584647836 
BIB |a Y 
SER |a journal 
JSO |a {"name":{"displayForm":["Lasse Elsemüller, Martin Schnuerch, Paul-Christian Bürkner, Stefan T. Radev"]},"id":{"eki":["1903735750"],"doi":["10.1037/met0000645"]},"person":[{"family":"Elsemüller","role":"aut","given":"Lasse","display":"Elsemüller, Lasse"},{"display":"Schnürch, Martin","family":"Schnürch","given":"Martin","role":"aut"},{"display":"Bürkner, Paul-Christian","given":"Paul-Christian","role":"aut","family":"Bürkner"},{"family":"Radev","given":"Stefan","role":"aut","display":"Radev, Stefan"}],"relHost":[{"id":{"issn":["1939-1463"],"eki":["363738002"],"zdb":["2103345-6"]},"part":{"pages":"?","year":"2024","text":"(2024), Seite ?","extent":"?"},"corporate":[{"role":"isb","display":"American Psychological Association"}],"pubHistory":["1.1996 -"],"language":["eng"],"type":{"media":"Online-Ressource","bibl":"periodical"},"note":["Gesehen am 11.02.05"],"title":[{"subtitle":"published quarterly by the American Psychological Association","title_sort":"Psychological methods","title":"Psychological methods"}],"origin":[{"dateIssuedDisp":"1996-","dateIssuedKey":"1996","publisherPlace":"Washington, DC ; [Erscheinungsort nicht ermittelbar]","publisher":"American Psychological Association ; Ovid"}],"disp":"A deep learning method for comparing Bayesian hierarchical modelsPsychological methods","physDesc":[{"extent":"Online-Ressource"}],"recId":"363738002"}],"origin":[{"dateIssuedDisp":"2024","dateIssuedKey":"2024"}],"title":[{"title":"A deep learning method for comparing Bayesian hierarchical models","title_sort":"deep learning method for comparing Bayesian hierarchical models"}],"note":["Gesehen am 30.09.2024"],"type":{"bibl":"article-journal","media":"Online-Ressource"},"language":["eng"],"recId":"1903735750","physDesc":[{"extent":"? S."}]} 
SRT |a ELSEMUELLEDEEPLEARNI2024 

Similar Items