Anisotropic Trust Region Warping for Prior-Data Fitted Networks in Bayesian Optimisation
Prior-Data Fitted Networks (PFNs) replace the hyperparameter refitting of Gaussian Process (GP) surrogates in Bayesian optimisation with a single forward pass, at the cost of a prior that is fixed at…
- 用途
- 原典で確認
- 難易度の目安
- Hard
- 実装コストの目安
- High
概要
Prior-Data Fitted Networks (PFNs) replace the hyperparameter refitting of Gaussian Process (GP) surrogates in Bayesian optimisation with a single forward pass, at the cost of a prior that is fixed at meta-training time. We show empirically that a PFN's predictive log-likelihood falls increasingly behind that of a refitted GP as the objective function's lengthscale grows beyond the mode of the PFN's meta-training prior, and that this deficit carries over to optimisation performance. To correct the mismatch without retraining the network, we introduce Anisotropic Trust Region (ATR) warping. Per-dimension lengthscales are estimated by kernel target alignment, and an affine rescaling of the input space maps the objective onto the lengthscale the PFN expects, with a trust region around the anchor determining which observations are retained, leaving the PFN's $\mathcal{O}(d \cdot N^2)$ inference cost unchanged. On functions drawn from a Matérn-3/2 kernel, ATR warping improves optimisation performance of the PFN at both extremes of the lengthscale range. Our implementation is publicly available at https://github.com/youngers2006/Trust-Region-Warping-For-PFNs.
参考リンク
リンク先のライセンス・再現性・データ条件をご確認ください。自動分類や実装の目安は保証ではありません。
実装を始めるプロンプト
参考リンクを確認したうえで、次の内容をコピーして利用できます。
# 論文・記事 Anisotropic Trust Region Warping for Prior-Data Fitted Networks in Bayesian Optimisation # 概要 Prior-Data Fitted Networks (PFNs) replace the hyperparameter refitting of Gaussian Process (GP) surrogates in Bayesian optimisation with a single forward pass, at the cost of a prior that is fixed at… # 原典 https://arxiv.org/abs/2610.37737 上記の手法を実装してください。実行前に依存関係とライセンスを確認してください。