
Succinct Non-Interactive Arguments for Batch NP (BARGs) from LWE, Part 1
Zhengzhong Jin teaches this session of MIT's 6.5630 Advanced Topics in Cryptography, continuing the course's treatment of succinct non-interactive arguments for batch NP, known as BARGs, built from the learning with errors (LWE) assumption. The lecture develops the cryptographic machinery needed for this construction, including locally somewhere statistically binding (SSB) commitments with local opening and probabilistically checkable proofs (PCPs) with fast online verification. Jin works through the technical definitions and building blocks at the chalkboard, setting up how these pieces combine to produce succinct arguments for batching many NP statements into one proof. As part one of a two-part treatment, the session focuses on laying the theoretical groundwork rather than completing the full construction. The lecture runs eighty-two minutes and assumes familiarity with prior course material on NP relations, lattice-based cryptography, and proof systems, making it suited to viewers already grounded in theoretical cryptography.