一句话：提出基于扩散模型的合成内核执行跟踪生成框架，用约束引导修复保证系统不变量，为下游机器学习诊断任务提供低成本数据增强。
工业系统中用于系统诊断的机器学习模型依赖内核执行跟踪来捕捉细粒度行为，但采集生产环境跟踪面临运行时开销、存储成本与隐私约束，真实数据十分稀缺。TraceSynth 将跟踪建模为多通道序列（事件类型、时间戳、CPU 亲和性、线程标识与进程元数据），采用基于 Transformer 的去噪扩散过程，并引入约束引导修复以强制满足系统不变量。作者在六个基准上进行评估，发现生成质量强烈依赖负载类型：在确定性、计算密集型负载（scimark2）上，上下文长度 L=4096 时合成增强达到 87.2% F1-Macro，仅比纯真实数据基线低 2.6 个百分点；上下文长度是最主要的质量因素，L=4096 相对 L=256 提升 104%，约束引导修复最多提升合成数据质量 4.3%。消融实验表明，轻量 2 通道模型以约一半计算成本保留了完整 6 通道模型 97–99% 的性能。该工作支持在生产可观测性流水线中低成本扩充内核执行跟踪，并帮助判断合成数据何时可替代有限的真实跟踪。
Machine learning models for system diagnostics rely on kernel execution traces to capture fine-grained system behavior, but collecting production traces in industrial systems is costly due to runtime overhead, storage demands, and privacy constraints. We present TraceSynth, a diffusion-based framework for generating synthetic kernel traces that augment limited real data for downstream ML tasks. TraceSynth models traces as multi-channel sequences (event types, timestamps, CPU affinity, thread identifiers, and process metadata) using a Transformer-based denoising diffusion process with constraint-guided repair to enforce system invariants. Across six benchmarks, results show strong workload dependence. For deterministic, compute-heavy workloads (scimark2), synthetic augmentation achieves 87.2% F1-Macro at context length L = 4096, only 2.6 percentage points below real-only baselines. Context length is the dominant quality factor, with L = 4096 yielding a + 104% relative improvement over L = 256, while constraint-guided repair improves synthetic data quality by up to 4.3%. Ablation studies show that lightweight 2-channel models retain 97–99% of the performance of full 6-channel models at roughly half the computational cost. TraceSynth supports cost-effective augmentation of kernel execution traces in production observability pipelines and helps identify when synthetic data can substitute for limited real traces.
DOI 原文 · arXiv · PDF（开放获取） · DBLP
@inproceedings{SehgalPPET26,
title = {TraceSynth: Generating Production-Quality Kernel Traces with Constraint-Guided Diffusion Models},
author = {Yuvraj Sehgal and Sneh Patel and Mahsa Panahandeh and Naser Ezzati-Jivan and Francois Tetreault},
booktitle = {FSE},
year = {2026},
doi = {10.1145/3803437.3805222}
}