For more explanation on the assumptions, please see RSA assumptions.
To improve the cryptanalysis of new RSA assumptions needed for Verifiable Delay Functions (VDFs), the Ethereum Foundation announced the following bounties at Real World Crypto 2020:
A bounty of $10,000 will be given for:
- An algorithm that solves the Adaptive Root Problem in an RSA group asymptotically faster than the fastest known algorithm for factoring an RSA number.1
- Reducing the Adaptive Root Assumption to one of these assumptions: Strong RSA Assumption, RSA Assumption, Diffie-Hellman Assumption, or proving the Adaptive Root Assumption is (non-)equivalent to Factoring in the Generic Ring Model.
For more details on the assumptions, see here: RSA Assumptions.
Bounties for lower bounds on modular squaring
The RSA VDF depends critically on the sequentiality of modular squaring. To this end the paper by Williams and Wesolowski explored lower bounds in modular squaring, and announced the following bounties for improving them (please refer to the paper for details and definitions):
- Prove that for all , SUM on -bit inputs requires depth at least for some . (That is, improve upon Krapchenko's lower bound for SUM)
- Prove that for all , SUM on -bit inputs requires depth at least . (That is, prove the "reasonable hypothesis" stated immediately after Theorem 2 of the paper. This bounty is in addition to the bounty above.)
- Prove that there is a such that for all , SUM has circuits of depth at most . (That is, refute the "reasonable hypothesis", and do so for all large enough input lengths .)
- Improve the average-case depth lower bound Theorem 3 from the paper to for some , for any algorithm computing MS-MOD2 on at least of the inputs.
Bounties for solving the RSA adaptive root problem in concrete RSA groups have been instantiated on an Ethereum smart contract:
For more details, see concrete instance bounties.
At this point in time, the complexity of the General Number Field Sieve is . ↩