Guofu Liao (Shenzhen University), Taotao Wang (Shenzhen University), Shengli Zhang (Shenzhen University), Jiqun Zhang (Shenzhen University), Long Shi (Nanjing University of Science and Technology), Dacheng Tao (Nanyang Technological University)

Fine-tuning large language models (LLMs) is crucial for adapting them to specific tasks, yet it remains computationally demanding and raises concerns about correctness and privacy, particularly in untrusted environments. Although parameter-efficient methods like Low-Rank Adaptation (LoRA) significantly reduce resource requirements, ensuring the security and verifiability of fine-tuning under zero-knowledge constraints remains an unresolved challenge. To address this, we introduce VeriLoRA, the first framework to integrate LoRA fine-tuning with zero-knowledge proofs (ZKPs), achieving provable security and correctness. VeriLoRA employs advanced cryptographic techniques---such as lookup arguments, sumcheck protocols, and polynomial commitments---to verify both arithmetic and non-arithmetic operations in Transformer-based architectures. The framework provides end-to-end verifiability for forward propagation, backward propagation, and parameter updates during LoRA fine-tuning, while safeguarding the privacy of model parameters and training data. Leveraging GPU-based implementations, VeriLoRA demonstrates practicality and efficiency through experimental validation on open-source LLMs like LLaMA, scaling up to 13 billion parameters. By combining parameter-efficient fine-tuning with ZKPs, VeriLoRA bridges a critical gap, enabling secure and trustworthy deployment of LLMs in sensitive or untrusted environments.

View More Papers

Lightweight Internet Bandwidth Allocation and Isolation with Fractional Fair...

Marc Wyss (ETH Zurich), Yih-Chun Hu (University of Illinois at Urbana-Champaign), Vincent Lenders (University of Luxembourg), Roland Meier (armasuisse), Adrian Perrig (ETH Zurich)

Read More

Looma: A Low-Latency PQTLS Authentication Architecture for Cloud Applications

Xinshu Ma (University of Edinburgh), Michio Honda (University of Edinburgh)

Read More

One Small Patch for a File, One Giant Leap...

Julian Rederlechner, Ulysse Planta, Ali Abbasi (CISPA Helmholtz Center for Information Security)

Read More