-
Hello !
The bandwidth weights received from the consensuses don't take into account the bandwidth we add to the simulated network when we inject adversarial nodes. I think that this could have bad co…
-
> Also, what service is actually intended to host the SIV server? If it’s a third party (e.g. aws), it has control over the election as well.
> [...]
> The goal is not to cast aspersions on the vend…
-
> Google’s tag manager is loaded on every page, and can entirely control the election
>
> This JS is just loaded on every page and can therefore entirely control the DOM.
>
> [...]
>
> The go…
-
Looking at the code, I noticed that stream IDs are hashed using the identity hash function (added in #1149).
Stream IDs are unique u64s, but I believe they are not necessarily evenly distributed ac…
-
Hello, I would like to know how the backdoor accuracy can reach 100% without defenses in the case of a semantic backdoor. When I perform a single attack by a single adversary after the model has conve…
-
0x52
high
# Adversary can sandwich oracle updates to exploit vault
## Summary
BLVaultLido added a mechanism to siphon off all wstETH obtained from mismatched pool and oracle prices. This was imple…
-
# Lines of code
https://github.com/code-423n4/2024-05-arbitrum-foundation/blob/6f861c85b281a29f04daacfe17a2099d7dad5f8f/src/rollup/RollupAdminLogic.sol#L145
https://github.com/code-423n4/2024-05-arbi…
-
The dapol security & privacy definitions rely on the tree being privately held by the custodian. It may be useful to be able to share the whole tree with a 3rd party. So it would then be necessary to …
-
Even with "is_poison"=false, the accuracy is only about 10%. When "is_poison"=true and batch_size=264, I get the results as follows:
When there are adversaries, accuracy of backdoor is about 100%,ac…
-
In "Unfetter Discover" section, the word "security" is misspelled.