Research map / Image generation / Autoregressive and latent generation
RIDER: 3D RNA Inverse Design with Reinforcement Learning-Guided Diffusion
Tianmeng Hu, Yongzheng Cui, Biao Luo, Ke Li
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add RL fine-tuning on 3D structural-similarity reward to the pretrained diffusion model (ablation) | RIDE · RNA 3D inverse design test set | GDT_TS: 0.33 -> 0.62 | Improved |
| add RL fine-tuning on structural reward (ablation) | RIDE · RNA 3D inverse design test set | RMSD (A): 10.36 -> 3.35 | Improved |
| add RL fine-tuning on structural reward (ablation) | RIDE · RNA 3D inverse design test set | native sequence recovery: 0.61 -> 0.45-0.49 | Hurt |
| use combined GDT_TS+RMSD reward instead of RMSD-only or GDT-only (ablation) | RIDER · RNA 3D inverse design test set | RMSD (A): 4.09 -> 3.35 (vs R_rmsd); 4.60 -> 3.35 (vs R_gdt) | Improved |
| use TM-score reward instead of RMSD reward (ablation) | RIDER · RNA 3D inverse design test set | GDT_TS / RMSD: GDT_TS 0.56 -> 0.62; RMSD 4.09 -> 4.31 | Mixed |
| add moving-average baseline to batch-mean advantage (ablation) | RIDER · RNA 3D inverse design test set | GDT_TS: about 15% improvement over RIDER-ADV | Improved |
| use advantage estimation instead of raw rewards (ablation) | RIDER · RNA 3D inverse design test set | GDT_TS / RMSD | Improved |
| replace RhoFold oracle with AlphaFold3 for reward (ablation) | RIDER · RNA 3D inverse design test set | GDT_TS: 0.62 -> 0.57 vs gRNAde 0.26 | No clear effect |
| use RIDE instead of RiboDiffusion and gRNAde | RIDE · RNA 3D inverse design test set | native sequence recovery: 0.56 -> 0.61 | Improved |
| increase native sequence recovery | RIDE designs · RNA 3D inverse design test set | GDT_TS | No clear effect |
| use AdamW with lr 5e-4 instead of Adam lr 3e-4 (ablation) | RIDE · RIDE hyperparameter sweep | NSR: 0.61 -> 0.47 | Hurt |
Closest papers by meaning
- gRNAde: Geometric Deep Learning for 3D RNA inverse design (ICLR 2025 · similarity 0.91)
- TANGO: RNA Topology and Geometry Co-Design (NeurIPS 2026 · similarity 0.90)
- RNAFlow: RNA Structure & Sequence Design via Inverse Folding-Based Flow Matching (ICML 2024 · similarity 0.90)
- RDesign: Hierarchical Data-efficient Representation Learning for Tertiary Structure-based RNA Design (ICLR 2024 · similarity 0.87)
- Beyond Sequence: Impact of Geometric Context for RNA Property Prediction (ICLR 2025 · similarity 0.86)
- BindGPT: A Scalable Framework for 3D Molecular Design via Language Modeling and Reinforcement Learning (AAAI 2025 · 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.
