[论文] ONAM: Fitting Orthogonal Neural Additive Models for Post-Hoc Interpret...
论文概要
研究领域: ML
作者: David Köhler, Matthias Schmid
发布时间: 2026-11-16
arXiv: 2611.21550
中文摘要
网络靶场和夺旗场景的构建仍然需要大量人力,因为作者必须协调基础设施、易受攻击的服务、攻击路径、工件和成功条件。LLM 可以提出详细的演练方案,但看似合理的描述并不能保证软件可部署或目标可验证。我们提出 ScenarioForge,将高级需求转化为有资源支撑的网络拓扑、服务、流量、分段和挑战链,然后在 Common Open Research Emulator(CORE)中实例化并验证它们。ScenarioForge 执行数据集包含 226 个物化的场景,涵盖 294 个 Vulhub 镜像和我们用 AI 辅助搭建开发的全部 85 个自定义挑战资源生成器。所有 226 个场景都通过了构建、预览、排序、部署、严格运行时验证和实时工件检查;它们的 867 个步骤(每个都是单独可验证的挑战目标)包含 457 个漏洞支撑和 410 个生成资源支撑的实例。在一个独立的 50 提示匹配基准测试中,ScenarioForge+Qwen3.8 和 ScenarioForge+GPT5.6 各实现了 50/50(100%)的拓扑/构建成功率和 160/160 的意图字段;而直接的前沿生成仅在完整累积脚手架(结构、词汇、策略和容器镜像指导)下达到 41/50 的构建和 147/160 的字段。
原文摘要
Modern machine learning models, such as artificial neural networks, achieve high predictive performance through increasingly complex black-box architectures, limiting their use in settings where transparency, interpretability, and fairness are required. Here we present the R package ONAM, which provides a novel and practical implementation of functional decomposition, a technique to facilitate the interpretability of black-box models through structured effect representations. ONAM implements an algorithmic framework for the post-hoc decomposition of predictions into components associated with different input features. The package is designed for usability and efficiency, including functionalities for visualization as well as quantitative measures to assess interpretability. As a model-agno...
自动采集于 2026-11-18
#论文 #arXiv #ML #小凯