AI Paper Atlas

Research map / Probabilistic and scientific modeling / Variational and generative sampling

VaiPhy: a Variational Inference Based Algorithm for Phylogeny

Hazal Koptagel, Oskar Kviman, Harald Melin, Negar Safinianaini, Jens Lagergren

NeurIPS 2022 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Annealing and divergence-based variational objectives · 4 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use phi-CSMC (VaiPhy-learned proposals) instead of vanilla CSMCCSMC · DS1log-likelihood: -8306.76 -> -7290.36Improved
use phi-CSMC instead of vanilla CSMCCSMC · DS3log-likelihood: -35381.01 -> -33798.06Improved
use phi-CSMC instead of vanilla CSMCCSMC · DS2log-likelihood: -27884.37 -> -30568.49Hurt
use phi-CSMC instead of VCSMCCSMC · DS1log-likelihood: -9180.34 -> -7290.36Improved
use phi-CSMC instead of VCSMCCSMC · DS6log-likelihood: -9296.66 -> -7013.83Improved
use VaiPhy instead of VCSMCphylogenetic inference · DS4log-likelihood: -17106.10 -> -13700.86Improved
use VaiPhy instead of vanilla CSMCphylogenetic inference · DS2log-likelihood: -27884.37 -> -31203.44Hurt
use MrBayes SS instead of VaiPhyphylogenetic inference · DS1log-likelihood: -7490.54 -> -7032.45Improved
use VBPI-NF instead of phi-CSMCphylogenetic inference · DS7log-likelihood: -9209.18 -> -8650.49Improved
use VaiPhy instead of VBPI-NF on single corewall-clock time · DS1runtime: 100 minutes -> 83 minutesImproved
use VaiPhy instead of VBPI-NFparameter count · DS1learnable parameters: 8.4 million -> 0.8 millionImproved
use VaiPhy instead of VBPI-NF to reach same accuracywall-clock time · DS1runtime to VaiPhy accuracy: 11xImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy