Proof aggregation compresses multiple proofs into one, revolutionizing proving frameworks. In this talk, I’ll show how this concept generalizes to sharding programs into smaller pieces, enabling efficient proof generation and verification in zkVMs.

In this presentation, we will introduce zkBugs (https://zkbugs.com/), a database containing vulnerabilities in ZK protocols. We will further present how we have leveraged zkBugs in our research projects, but also how we can learn from past bugs.

This talk introduces Pico, a next-generation zkVM by Brevis that is modular, performant, and future-proof. It enables developers to build ZK-powered dApps using plain Rust, leveraging multiple proving backends for optimal efficiency and flexibility.

Testing zkEVM provers in production is critical, yet best practices are lacking. We introduce a framework using historical and synthetic traces to enhance robustness and optimize deployment, ensuring zkEVM constraints are validated with confidence.

An analysis of how to prove the Ethereum Execution Layer in under a slot time. We'll explore how to gain precious seconds on each step of the proving workflow, from preparing prover inputs and generating a trace to generating the STARK.

I explore differences between commitment- and encryption-based solutions for confidential payment. I focus on the features an encryption-based solution could enable, like composability and the flexible programming of viewing and transferring assets.

This talk examines challenges in signing encrypted transactions with FHE for MEV mitigation, proposing solutions like Account Abstraction, Chameleon hashes, and online/offline signatures to replace the less FHE-friendly keccak and ECDSA.

Explore ZK coprocessors for cross-chain data verification & modular on-chain compute, diving into design of ZK coprocessor, use cases (solvency proofs, compliance), and live demo of a bridge solvency check module, verify its STARK proof on-chain.

We document our research into using AI to identify vulnerabilities in zero-knowledge proof systems. We share insights from experimenting with different prompting techniques—from basic methods to advanced multi-agent and chain-of-thought approaches

We introduce OpenVM, a performant and modular zkVM framework built for customization and extensibility. OpenVM allows developers to easily use ZK without trading off performance. We will discuss the core OpenVM design and applications enabled by it.

Lookup arguments are commonly used in proof generation by Layer 2 solutions. They assert that the elements of one vector are a subset of another vector. We focus on a special case where a bilateral order-preserving lookup is required.

Threshold encrypted mempools (TEMs) can prevent malicious MEV extraction but crucially rely on a trust assumption, which hinders adoption in practice. This talk presents a novel solution to reduce this assumption, enabling widespread use of TEMs.

We introduce a simple custom PLONK-ish based SNARK and we describe how it can be used to efficiently aggregate BLS signatures which are prevalent in many consensus mechanisms nowadays. We also show how to prove its security which requires O-SNARKs.

Currently the barriers to running a full node is 2TB of storage and most users use centralized JSON-RPC providers such as Infura. What if there was a way to reduce the requirements to partipate for full nodes utilizing EIP-4444's + Portal.