发表机构
Institute of Automation, Chinese Academy of Sciences; State Key Laboratory of Multimodal Artificial Intelligence Systems（中国科学院自动化研究所; 多模态人工智能系统国家重点实验室）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究外包变压器推理问题，提出GKR-HND协议验证同态-非同态分解变压器多项式主干，保留验证器检查记录与开口，委托公共评估给计算工作者，实验验证了证明路径与委托计算，无需密集矩阵重放。
AI 中文摘要
外包变压器推理使客户端面临模型替换和执行不完整的问题，而直接重放消除了委托的计算优势。我们提出了GKR-HND，这是一种用于验证同态-非同态分解变压器多项式主干的注册模型协议。保留的验证器检查GKR记录和注册权重开口，但将昂贵的公共评估委托给指定的计算工作者。假设保留的验证器诚实且证明者-工作者不勾结，验证器仅在工作者的签名、请求绑定响应与证明声明一致时才接受。使用预训练的HND模型进行的实验验证了证明路径和委托的公共计算，无需密集矩阵重放。
英文摘要
Outsourced Transformer inference exposes clients to model substitution and incomplete execution, while direct replay removes the computational benefit of delegation. We present GKR-HND, a registered-model protocol for verifying the polynomial backbone of Homomorphic--Nonhomomorphic Decomposition Transformers. The retained verifier checks the GKR transcript and registered-weight openings, but delegates expensive public evaluations to an assigned computation worker. Assuming an honest retained verifier and prover--worker non-collusion, the verifier accepts only when the worker's signed, request-bound response agrees with the proof claims. Experiments with pretrained HND models validate the proof path and the delegated public computation without dense-matrix replay.
Comments24 pages, including 4 pages of supporting information; 2 figures
