提出DiffGBM，用树模型实现高效精准的表格回归概率建模。
Conditioning Tree-Based Diffusions and Flows for Probabilistic Tabular Regression
- 基于高斯路径流匹配，直接学习速度场，支持快速确定性采样。
- 在11个基准上全胜基线，最优CRPS达0.725，显著提升预测精度。
- 适合追求高精度与可解释性的表格数据建模者使用。
基于树的扩散模型可在不使用神经密度估计器的情况下灵活拟合表格回归的条件预测分布，但其设计默认（如加噪路径、参数化、训练分布、特征、采样器）继承自神经设置，成为关键瓶颈：梯度提升集成实际解决的是由其条件决定的监督回归问题。本文提出DiffGBM，沿两个维度显式化这些设计。第一，采用高斯路径流匹配训练器 $p(y igm| x)$，直接学习速度场并代数恢复得分，支持少步确定性微分方程采样。第二，将得分侧策略——残差化、EDM风格预处理、对数标准差时间采样、噪声水平特征、损失加权、直方图分辨率——视为共享LightGBM表面的可调轴，而非固定组合。该「score-flex」空间将已有方法作为特例；在十一项表格基准上，折痕0调参、折痕1–5评估，40次试验预算与相同采样器下，所选配置在所有数据集上均优于基线（配对威尔科克森检验 $11/0$，$p<10^{-3}$），最优综合CRPS为0.725（基线0.699），超越任何同行工作。两行互补：score-flex以随机采样换取精度，最慢；流匹配采样快5.2倍，且校准最佳。经调优的非扩散基线仍能独占部分数据集，且随机采样器（$\>0$）未在帕累托意义上超越确定性解。
原文摘要 · Abstract (English)
Tree-based diffusion models fit flexible conditional predictive distributions for tabular regression without a neural density estimator, but they inherit their design defaults---noising path, parameterization, training distribution, features, sampler---from the neural setting. We show these defaults are the binding constraint: what a gradient-boosted ensemble actually solves is a supervised regression problem whose conditioning they determine. We present DiffGBM, which makes them explicit along two axes. First, a Gaussian-path flow-matching trainer for $p(y \mid x)$ that learns a velocity field directly and recovers the score algebraically, admitting few-step deterministic ODE sampling. Second, we expose the score-side recipe---residualization, EDM-style preconditioning, log-sigma time sampling, noise-level features, loss weighting, and histogram resolution---as jointly tunable axes over a shared LightGBM surface rather than one frozen bundle. This \emph{score-flex} space represents the published recipe as a special case; across eleven tabular benchmarks under fold-0 tuning, folds-1--5 evaluation, and a matched 40-trial budget and sampler, the selected configurations beat that baseline on \emph{every} dataset (paired Wilcoxon $11/0$, $p<10^{-3}$), with the best aggregate CRPS skill (0.725 vs.\ 0.699) of any row. The two rows are complementary: score-flex buys accuracy with a stochastic sampler and is the slowest row, while flow matching is the cheapest sampler ($5.2\times$ faster than the published baseline) and the best-calibrated DiffGBM row. Tuned non-diffusion baselines still win individual datasets, and stochastic ($\varepsilon>0$) flow samplers do not Pareto-dominate the deterministic corner.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。