Research map / Probabilistic and scientific modeling / Bayesian optimization
Symmetry-Aware Bayesian Optimization via Max Kernels
Anthony Bardou, Antoine Gonon, Aryan Ahadinia, Patrick Thiran
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 |
|---|---|---|---|
| replace orbit-averaged kernel with PSD-projected max kernel (ablation) | GP-UCB · Rastrigin5d (|G|=3840) | cumulative regret: 1583.5 -> 813.4 | Improved |
| replace orbit-averaged kernel with PSD-projected max kernel (ablation) | GP-UCB · Griewank6d (|G|=64) | cumulative regret: 3067.4 -> 1832.6 | Improved |
| replace orbit-averaged kernel with PSD-projected max kernel (ablation) | GP-UCB · Ackley2d (|G|=8) | cumulative regret: 128.2 -> 126.4 | No clear effect |
| replace orbit-averaged kernel with PSD-projected max kernel (ablation) | GP-UCB · Scaling2d (continuous group) | cumulative regret: 3361.8 -> 25.4 | Improved |
| replace orbit-averaged kernel with PSD-projected max kernel (ablation) | GP-UCB · Radial2d (continuous group) | cumulative regret: 480.9 -> 199.7 | Improved |
| replace orbit-averaged kernel with PSD-projected max kernel (ablation) | GP-UCB · WLAN8d | negated simple reward: -51.8 -> -74.4 | Improved |
| add orbit averaging over a continuous group vs base kernel (ablation) | GP-UCB · Radial2d (continuous group) | cumulative regret: 388.6 -> 480.9 | Hurt |
| add orbit averaging over a continuous group vs base kernel (ablation) | GP-UCB · Scaling2d (continuous group) | cumulative regret: 1820.6 -> 3361.8 | Hurt |
| increase symmetry group size (hyperoctahedral, d=1..5) (ablation) | orbit-averaged kernel · Ackley/Rastrigin d-dim | regret after 50 iterations | Hurt |
| use orbit averaging instead of the base kernel when |G| is moderate | GP-UCB · Ackley2d (|G|=8) | cumulative regret: 382.7 -> 128.2 | Improved |
Closest papers by meaning
- Sample-efficient Bayesian Optimisation Using Known Invariances (NeurIPS 2024 · similarity 0.87)
- Omnipresent Yet Overlooked: Heat Kernels in Combinatorial Bayesian Optimization (NeurIPS 2025 · similarity 0.87)
- BOIDS: High-Dimensional Bayesian Optimization via Incumbent-Guided Direction Lines and Subspace Embeddings (AAAI 2025 · similarity 0.87)
- Automated Kernel Discovery Towards Understanding High-dimensional Bayesian Optimization (NeurIPS 2026 · similarity 0.87)
- A General Framework for User-Guided Bayesian Optimization (ICLR 2024 · similarity 0.86)
- Adaptive Kernel Design for Bayesian Optimization Is a Piece of CAKE with LLMs (NeurIPS 2025 · similarity 0.86)
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.
