Research map / Probabilistic and scientific modeling / Bayesian optimization
Cliqueformer: Model-Based Optimization with Structured Transformers
Jakub Grudzien Kuba, Pieter Abbeel, Sergey Levine
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 FGM clique decomposition and clique-wise VIB on a transformer (ablation) | Cliqueformer · Superconductor | normalized top-1% score: 0.96 -> 1.43 | Improved |
| add FGM clique decomposition and clique-wise VIB on a transformer (ablation) | Cliqueformer · Lat. RBF 31 | normalized top-1% score: 0.00 -> 0.64 | Improved |
| add FGM clique decomposition and clique-wise VIB on a transformer (ablation) | Cliqueformer · DNA HEPG2 | normalized top-1% score: 2.13 -> 2.10 | No clear effect |
| use Cliqueformer vs. COMs on low-dim latent RBF | Cliqueformer · Lat. RBF 11 | normalized top-1% score: 0.66 (COMs) -> 0.65 | No clear effect |
| use Cliqueformer vs. COMs as latent dimension grows | Cliqueformer · Lat. RBF 61 | normalized top-1% score: 0.25 -> 0.66 | Improved |
| use Cliqueformer vs. MatchOpt | Cliqueformer · Lat. RBF 61 | normalized top-1% score: 0.74 -> 0.66 | Hurt |
| use Cliqueformer vs. best baseline overall | Cliqueformer · 7-task average | average score: 1.07 (MatchOpt) -> 1.36 | Improved |
| use a single clique (Nclique = 1) (ablation) | Cliqueformer · TFBind-8, Superconductor | design score | Hurt |
| use too many cliques with fixed latent size (ablation) | Cliqueformer · TFBind-8, Superconductor | design score | Hurt |
Closest papers by meaning
- Offline Materials Optimization with CliqueFlowmer (NeurIPS 2026 · similarity 0.88)
- Aligning Transformers with Continuous Feedback via Energy Rank Alignment (NeurIPS 2025 · similarity 0.84)
- Offline Model-Based Optimization via Policy-Guided Gradient Search (AAAI 2024 · similarity 0.84)
- Deep Generative Model for Mechanical System Configuration Design (AAAI 2025 · similarity 0.84)
- Robust Model-Based Optimization for Challenging Fitness Landscapes (ICLR 2024 · similarity 0.84)
- Amortized Molecular Optimization via Group Relative Policy Optimization (NeurIPS 2026 · similarity 0.83)
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.
