You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The main challenge is understanding how these implementations use a univariate polynomial IOP (STARK) to integrate GKR, which requires a multivariate polynomial IOP.
Miden VM provides a clear approach to this integration (detailed in [this doc] and [this doc]). However, its GKR and related APIs are from Winterfell (logup-gkr branch), making direct application to the Stwo backend difficult. - This may not be a blocker, as Stwo already implements GKR and sumcheck, meaning integration could be possible with modifications.
Stwo's GKR implementation is clear for proving a logup lookup alone, but its integration with STARK is less clear.
One key aspect is understanding the "univariate IOP for multilinear evaluation at a point" in this example, which seems central to the integration.
My current research is focused on understanding this component.
Next Steps
I will continue investigating this and update the issue as I gain a clearer picture. Any insights or related references are welcome!
The text was updated successfully, but these errors were encountered:
Logup-GKR Integration in STARK Provers
Context
I have been researching how different implementations integrate Logup-GKR with STARK proofs. My focus is on two implementations:
Current Progress
The main challenge is understanding how these implementations use a univariate polynomial IOP (STARK) to integrate GKR, which requires a multivariate polynomial IOP.
Miden VM provides a clear approach to this integration (detailed in [this doc] and [this doc]). However, its GKR and related APIs are from Winterfell (logup-gkr branch), making direct application to the Stwo backend difficult. - This may not be a blocker, as Stwo already implements GKR and sumcheck, meaning integration could be possible with modifications.
Stwo's GKR implementation is clear for proving a logup lookup alone, but its integration with STARK is less clear.
Next Steps
I will continue investigating this and update the issue as I gain a clearer picture. Any insights or related references are welcome!
The text was updated successfully, but these errors were encountered: