Verde: Verification via Refereed Delegation for Machine Learning Programs
A. Arun, A. St. Arnaud, A. Titov, B. Wilcox, V. Kolobaric, M. Brinkmann, O. Ersoy, B. Fielding, J. Bonneau. 2025. arXiv.
See what cites it| Link | https://arxiv.org/abs/2502.19405 |
|---|---|
| arXiv | 2502.19405 |
| Version | v1 (26 Feb 2025) read in HTML. dblp lists only the arXiv version. |
| Accessed | 2026-09-24 |
| Note | Most authors at Gensyn; Arun and Bonneau at New York University (Bonneau also a16z crypto research). Describes the Verde dispute protocol (correct result if at least one provider is honest) and RepOps; evaluation reports RepOps overheads on T4, RTX 3090 and A100 GPUs for DistilBERT and Llama models (Table 2: Llama-8B on A100 80 GB, 98% inference, 126% LoRA fine-tuning). RepOps supports FP32 and one GPU per setup. |