报告人：Alan Szepieniec
报告地点：淦昌苑D座104报告厅
报告时间：2025年11月27日/28日 14:00-16:00
报告人简介：
Alan Szepieniec is a cryptography researcher and entrepreneur based in Switzerland. After obtaining a PhD in post-quantum cryptography from KU Leuven, he founded cryptography research lab AS Discrete Mathematics, which did contract work for Nervos Foundation and Ethereum Foundation, among others. During this time he published important contributions to the field of interactive proof systems, such as DARK/Supersonic, which introduced transparent SNARKs based on groups of unknown order as well as the Polynomial IOP formalism, and such as the Marvellous family of arithmetization-oriented hash function whose most well-known member, Rescue, was selected by StarkWare's EthStark project. In 2021 he co-founded the Neptune Cash cryptocurrency project, and which launched its mainnet in February of 2025. This launch marked the first time zk-STARKs were natively present as a first-class citizen hardcoded into the consensus mechanism of a Layer-1 blockchain.
报告一：Introduction to STARKs
报告简介：
Zk-STARKs, defined narrowly as hash-based (non-)interactive proof systems based on DEEP-ALI and low-degree tests, are pervasive throughout the blockchain space and have many applications in other fields as well. These lectures cover the building blocks for their construction, from theory to practice.
A universal recipe for constructing SNARKs consists of two steps:
1) equip the verifier with superpowers that make verifying the prover's claim easy; and
2) strip away those superpowers by emulating them with cryptography.
Applied to zk-STARKs, this universal recipe provides the blue print for a compilation pipeline that puts DEEP-ALI at the top and a concrete non-interactive STARK at the bottom. These lectures follow that blueprint; motivate every intermediate stage and cover every compilation step with a discussion of the induced soundness error.
报告二：Advanced zk-STARKs
报告简介：
After the elaborate guided tour along the STARK compilation pipeline, the focus shifts to three rather more optional topics:
1) Optimizations -- techniques deployed in practice to improve concrete metrics.
2) Qualitative improvements -- some common features, such as zero-knowledge, among others, are not incremental but binary.
3) STARKs in the context of VM-based proof systems.
Finally, the focus turns to a selection topics not covered in any detail but which do nevertheless have practical and research significance.
The presentations for these lectures are available at these URLs:
- https://asz.ink/presentations/2025-09-18-Introduction-to-STARKs.pdf
- https://asz.ink/presentations/2025-09-18-Advanced-zkSTARKs.pdf
邀请人：张云聪
审核人：魏普文
