AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

SymSpectra: Symmetric Information Bottleneck Framework for Molecular Structure Recognition under Imbalanced Settings

Xiaohan Qin, Wenjie Du, Yang Wang

ICML 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Dense features and foundation priors

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 SymSpectra's SCIB + CMI weighting + LSTM decoder instead of the best baselineSymSpectra · Simulated spectra (IR, 13C-NMR, 1H-NMR)F1: 0.947 -> 0.970Improved
add focal loss to the Alberts et al. baselineAlberts et al. · Simulated spectraF1: 0.947 -> 0.956Improved
use CMI dynamic weighting instead of focal loss, oversampling or reweightingSymSpectra · Simulated spectraF1: 0.956 -> 0.970Improved
add CMI dynamic weighting (ablation)SymSpectra · low-frequency functional groupsF1: 0.844 -> 0.866Improved
add SCIB compression (ablation)SymSpectra · low-frequency functional groupsF1: 0.827 -> 0.866Improved
add LSTM decoder for label dependencies (ablation)SymSpectra · low-frequency functional groupsF1: 0.854 -> 0.866Improved
add LSTM decoder for label dependencies (ablation)SymSpectra · high-frequency functional groupsF1: 0.966 -> 0.967No clear effect
use sequential prediction with IUPAC order instead of non-sequential CNN (ablation)SymSpectra decoder · Imine functional groupF1: 0.540 -> 0.885Improved
use sequential prediction instead of non-sequential CNN (ablation)SymSpectra decoder · Carboxylic acid functional groupF1: 0.992 -> 0.974Hurt
use scaffold-disjoint test splitSymSpectra · structurally disjoint test setF1: 0.970 -> 0.8798Hurt
evaluate on experimental SDBS spectra instead of simulatedSymSpectra · multimodal spectraF1: 0.970 -> 0.919Hurt
set s1 = s2 = 0.3 and beta = 1e-6 (ablation)SymSpectra · hyperparameter sensitivityF1Improved

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