Research map / Probabilistic and scientific modeling / Molecules and proteins
Probabilistic Transformer: Modelling Ambiguities and Distributions for RNA Folding and Molecule Design
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| add kappa annealing, init 0.2 (ablation) | ProbTransformer · RNA folding | Hamming distance: 32.2 -> 29.2 | Helped |
| add kappa annealing, init 0.1 (ablation) | ProbTransformer · RNA folding | Hamming distance: 29.4 -> 27.9 | Helped |
| add kappa annealing, init 0.05 (ablation) | ProbTransformer · RNA folding | Hamming distance: 28.0 -> 28.0 | No clear effect |
| add kappa annealing, init 0.02 (ablation) | ProbTransformer · RNA folding | Hamming distance: 29.4 -> 29.2 | No clear effect |
| replace vanilla Transformer with ProbTransformer | Transformer · TS0 | F1 score: 50.5 -> 62.5 | Helped |
| replace vanilla Transformer with ProbTransformer | Transformer · TSsameStruc | Hamming distance: 4.6 -> 3.2 | Helped |
| replace vanilla Transformer with ProbTransformer | Transformer · TS0 | fraction solved: 0.084 -> 0.118 | Helped |
| replace softmax sampling with ProbTransformer latent sampling | Transformer · Synthetic task | KL divergence: 7.84 -> 0.52 | Helped |
| replace MC dropout sampling with ProbTransformer | Transformer · Synthetic task | validity: 0.93 -> 0.99 | Helped |
| replace MolGPT with ProbTransformer | MolGPT · GuacaMol | TPSA SD: 4.80 -> 2.04 | Helped |
| replace MolGPT with ProbTransformer | MolGPT · GuacaMol | uniqueness: 0.969 -> 0.821 | Hurt |
| replace MolGPT with ProbTransformer | MolGPT · GuacaMol | validity: 0.973 -> 0.981 | Helped |
Closest papers by meaning
- Molecule Design by Latent Prompt Transformer (NeurIPS 2024 · similarity 0.88)
- GeoMFormer: A General Architecture for Geometric Molecular Representation Learning (ICML 2024 · similarity 0.85)
- Generalist Equivariant Transformer Towards 3D Molecular Interaction Learning (ICML 2024 · similarity 0.85)
- Deciphering RNA Secondary Structure Prediction: A Probabilistic K-Rook Matching Perspective (ICML 2024 · similarity 0.85)
- Swallowing the Bitter Pill: Simplified Scalable Conformer Generation (ICML 2024 · similarity 0.85)
- RNAFlow: RNA Structure & Sequence Design via Inverse Folding-Based Flow Matching (ICML 2024 · similarity 0.85)
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.