AI Paper Atlas

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

A Diffusive Classification Loss for Learning Energy-based Generative Models

RuiKang OuYang, Louis Grenioux, José Miguel Hernández-Lobato

ICML 2026 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Score-based generative modeling analysis · 0 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
add the DiffCLF classification loss to DSMVP diffusion model on MOG-40 · MOG-40, dim 8 to 128classification loss: 383.53 -> 4.40 (dim 128, DSM -> DiffCLF)Improved
add the DiffCLF classification loss to DSMVP diffusion model on MOG-40 · MOG-40Fisher divergence and MMDNo clear effect
use DiffCLF instead of conditional time score matching (CtSM) as the auxiliary loss to DSM (ablation)VP diffusion model on MOG-40 · MOG-40, dim 8classification loss: 6.80 -> 4.41Improved
add DiffCLF to the stochastic interpolant energy trainingSI between MOG-40 and MOG-2 · d=2 and d=128 log-density fitlog-density agreement with ground truthImproved
train with DiffCLF in addition to DSM for molecular systemsenergy-based diffusion models · Muller-Brown, Alanine Dipeptide, ChignolinLangevin sample quality with learned energyImproved
use DiffCLF instead of Fokker-Planck regularization (ablation)energy-based diffusion models on molecules · molecular systemstraining speed: 1.5-2.5x faster training at comparable performanceImproved
use DiffCLF-trained energies instead of DSM for SMC compositionOR and AND toy composition · toy compositionscomposition accuracyImproved
use DiffCLF-trained energies instead of DSM or CtSM in SMC Boltzmann generatorSI MOG-40 to MOG-2, 512-step SMC · Sliced Wasserstein and KS distanceW2 and KS distanceImproved
add Lclf to the Lbase objective for neural thermodynamic integrationLbase (Mate et al., 2025) · ALDP solvation free energyfree energy estimate (reference 29.43): 27.30 -> 29.02Improved
increase number of classification levels N (ablation)DiffCLF · theory (Corollary D.9)asymptotic covariance normImproved

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