---
格式版本: 2
标题: "Leaf Values as Coordinates: Exact Contrastive Explanation for Gradient-Boosted Ensembles"
原文链接: "https://arxiv.org/abs/2608.19127"
发布日期: "2026-08-19"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Wed, 19 Aug 2026 17:20:33 UTC (41 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-08-20T15:51:57+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-08-20T15:48:16+08:00"
入库时间: "2026-08-20T07:51:57.741Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=model%20fit&searchtype=all"
匹配关键词:
  - "model fit"
  - "AI"
相关厂家:
  []
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 7
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "该资料为arXiv机器学习论文，主题为梯度提升树的可解释性，与超节点/AI Rack/机柜级AI基础设施等完全无关，无任何相关技术、商业或部署信息，判为无关内容。"
AI质检模型: "ali-deepseek-v4-flash"
AI质检时间: "2026-08-20T15:57:39+08:00"
AI主题相关性: 0
AI来源权威性: 2
AI新颖性: 0
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 5
AI摘要: "该文提出将梯度提升树的叶值视为坐标，使对比解释成为精确计算：两个实例的差异向量可逐项追溯至真实树分裂。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-07T03:19:36.802Z"
采集批次: "2026年8月20日14点19分32秒"
采集批次ID: "20260820-141932-079"
去重键: "https://arxiv.org/abs/2608.19127"
---

## Computer Science > Machine Learning

## Title:Leaf Values as Coordinates: Exact Contrastive Explanation for Gradient-Boosted Ensembles

Authors:[Emanuele Luzio](https://arxiv.org/search/cs?searchtype=author&query=Luzio,+E)

[View PDF](https://arxiv.org/pdf/2608.19127) [HTML (experimental)](https://arxiv.org/html/2608.19127v1)

> Abstract:A gradient-boosted ensemble predicts by summing one leaf value per tree. Read  
> those values as coordinates rather than as intermediate results, and every  
> instance becomes a point in R^M on which the model acts linearly: the score is  
> the sum of the coordinates.  
> This small change of view makes contrastive explanation exact. The difference  
> between two instances is a vector that is identically zero wherever they share  
> a leaf, so the gap between a rejected applicant and an accepted one is carried  
> by a handful of coordinates, each traceable to a real split in a real tree.  
> Nothing is fitted, sampled, or assumed additive in features -- the additivity  
> is already there, in the right space.  
> We build a recourse method on this representation and evaluate it on five  
> tabular datasets under repeated cross-validation. Its recommendation  
> reconstructs the model's own decision to 6.2 x 10^-15, so an auditor can  
> re-check the arithmetic without the model. On the credit datasets it is  
> Pareto-non-dominated on effort against realism. And when recommendations are  
> restricted to changes the subject could actually make -- not their age, not a  
> settled delinquency -- it retains 58% of its validity where the strongest  
> baseline retains 41%, a distinction the standard evaluation cannot see because  
> it never asks whether a recommendation can be carried out.

| Comments: |  |
| --- | --- |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computers and Society (cs.CY) |
| ACM classes: | I.2.6; K.4.1 |
| Cite as: | [arXiv:2608.19127](https://arxiv.org/abs/2608.19127) \[cs.LG\] |
|  | (or [arXiv:2608.19127v1](https://arxiv.org/abs/2608.19127v1) \[cs.LG\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.19127](https://doi.org/10.48550/arXiv.2608.19127) |

## Submission history

From: Emanuele Luzio \[[view email](https://arxiv.org/show-email/b90cc312/2608.19127)\]  
**\[v1\]** Wed, 19 Aug 2026 17:20:33 UTC (41 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.19127) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))