Automatic Differentiation is no Panacea for Phylogenetic Gradient Computation.
Bayesian inference
gradient
phylogenetics
variational inference
Journal
Genome biology and evolution
ISSN: 1759-6653
Titre abrégé: Genome Biol Evol
Pays: England
ID NLM: 101509707
Informations de publication
Date de publication:
01 06 2023
01 06 2023
Historique:
accepted:
25
05
2023
medline:
22
6
2023
pubmed:
2
6
2023
entrez:
2
6
2023
Statut:
ppublish
Résumé
Gradients of probabilistic model likelihoods with respect to their parameters are essential for modern computational statistics and machine learning. These calculations are readily available for arbitrary models via "automatic differentiation" implemented in general-purpose machine-learning libraries such as TensorFlow and PyTorch. Although these libraries are highly optimized, it is not clear if their general-purpose nature will limit their algorithmic complexity or implementation speed for the phylogenetic case compared to phylogenetics-specific code. In this paper, we compare six gradient implementations of the phylogenetic likelihood functions, in isolation and also as part of a variational inference procedure. We find that although automatic differentiation can scale approximately linearly in tree size, it is much slower than the carefully implemented gradient calculation for tree likelihood and ratio transformation operations. We conclude that a mixed approach combining phylogenetic libraries with machine learning libraries will provide the optimal combination of speed and model flexibility moving forward.
Identifiants
pubmed: 37265233
pii: 7188956
doi: 10.1093/gbe/evad099
pmc: PMC10282121
pii:
doi:
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Research Support, N.I.H., Extramural
Langues
eng
Sous-ensembles de citation
IM
Subventions
Organisme : NIH HHS
ID : S10 OD028685
Pays : United States
Organisme : Howard Hughes Medical Institute
Pays : United States
Informations de copyright
© The Author(s) 2023. Published by Oxford University Press on behalf of Society for Molecular Biology and Evolution.
Références
Virus Evol. 2018 Jun 08;4(1):vey016
pubmed: 29942656
Mol Biol Evol. 2022 Aug 3;39(8):
pubmed: 35816422
Virus Evol. 2018 Jan 08;4(1):vex042
pubmed: 29340210
Genome Res. 1998 Mar;8(3):222-33
pubmed: 9521926
Mol Biol Evol. 2020 Oct 1;37(10):3047-3060
pubmed: 32458974
Genome Res. 2021 Nov;31(11):2107-2119
pubmed: 34426513
Elife. 2014;3:e01914
pubmed: 24497547
Nat Biotechnol. 2017 Apr 11;35(4):316-319
pubmed: 28398311
Syst Biol. 2019 Nov 1;68(6):1052-1061
pubmed: 31034053
Syst Biol. 2021 Feb 10;70(2):258-267
pubmed: 32687171
Stat Appl Genet Mol Biol. 2012 Sep 25;11(4):Article 14
pubmed: 23023698
Mol Biol Evol. 2019 Apr 1;36(4):825-833
pubmed: 30715448
J Mol Evol. 1981;17(6):368-76
pubmed: 7288891
Syst Biol. 2020 Mar 1;69(2):209-220
pubmed: 31504998
BMC Evol Biol. 2014 Jul 24;14:163
pubmed: 25055743
J Stat Softw. 2017;76:
pubmed: 36568334
PeerJ. 2019 Dec 18;7:e8272
pubmed: 31976168