Research

Research seriesL3algorithms

Zero-Knowledge Proofs and Verifiable AI

A zero-knowledge proof lets one party convince another that a computation was performed correctly without revealing the inputs — which is exactly what verifiable AI inference needs. This series builds the idea from interactive proofs and the simulator argument through Fiat-Shamir, SNARKs, transparent STARKs, and closes with zkML: proving a model produced an output on a committed input. Grounded in Goldwasser-Micali-Rackoff, Groth16, and Ben-Sasson STARKs.

Murali Chillakuru·5 articles
  1. 1
    Interactive Proofs and Zero-Knowledge, Defined: Completeness, Soundness, and the Simulator

    A zero-knowledge proof convinces a verifier that a statement is true while revealing nothing else. The definition rests on three properties and one beautiful idea: the simulator.

  2. 2
    From Interactive to Non-Interactive: Fiat-Shamir, Commitments, and the Random-Oracle Caveat

    Live back-and-forth is inconvenient. The Fiat-Shamir transform turns an interactive proof into a single string anyone can check by letting a hash function play the verifier's role.

  3. 3
    SNARKs: Succinct Arguments, the Arithmetic-Circuit Pipeline, and the Trusted-Setup Question

    A SNARK proves a huge computation was done correctly with a proof so small and quick to check it seems impossible. The price is a compilation pipeline and, often, a trusted setup.

  4. 4
    STARKs and Transparency: Hash-Based, Post-Quantum-Plausible Proofs Without Trusted Setup

    STARKs drop the trusted setup entirely and rest on hash functions alone — which makes them transparent and plausibly quantum-safe, at the cost of larger proofs.

  5. 5
    Proving an ML Inference: zkML, the Committed-Input Guarantee, and the Honest Cost

    zkML applies zero-knowledge proofs to machine learning: a prover convinces anyone that a model produced a specific output on a committed input, without revealing the input or the weights.