arXiv：cs.LG· Souvik Das, Ariful Azad, Nitesh Saxena·· 10 小时前
基于图的私有 Transformer 推理可比性、规模与归因分析
A Graph-Based Analysis of Comparability, Scale, and Attribution in Private Transformer Inference
AI 导读
一项研究构建了包含 62 个私有 Transformer 推理（PTI）系统的经验图谱，将系统属性映射为节点特征、评估兼容性映射为边。节点级归因显示密码学进展驱动了多数系统的性能提升，但总与其他因素共同出现；边级审计显示 1891 个候选系统配对中仅 1 对在共享假设下可直接比较。图级分析显示私有推理的规模前沿已停滞在 13B 参数。
正文
Abstract:Rising privacy concerns around large language models have accelerated research in private Transformer inference (PTI). Synthesizing the empirical claims in this line of research, however, is hindered by multi-dimensional cryptographic progress and inconsistent evaluation contracts. In this work, we systematize the field by constructing an empirical graph of 62 PTI systems, mapping system attributes to node features and evaluation compatibility to edges. A three-tiered analysis of this graph illuminates the state of private inference. Node-level attribution shows that cryptographic progress drives performance gains across most systems, yet always co-occurs with other factors (model fidelity, hardware, cost accounting). Edge-level auditing shows that only 1 of 1891 candidate system pairs is directly comparable under shared assumptions. Finally, graph-level analysis shows that the scaling frontier of private inference has plateaued at 13B parameters and projects the cost of deploying current PTI techniques on production-scale Transformers. Overall, this graph-based framework equips the community to evaluate PTI claims transparently and channel technical speedups into deployment-scale private inference.
| Comments: | 21 pages, 10 figures, 11 tables |
| Subjects: | Cryptography and Security (cs.CR); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.19607 [cs.CR] |
| (or arXiv:2610.19607v1 [cs.CR] for this version) | |
| https://doi.org/10.48550/arXiv.2610.19607 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Souvik Das [view email]
[v1]
Wed, 14 Oct 2026 05:30:25 UTC (448 KB)
来源：arXiv：cs.LG · arxiv.org