-
I would like to ask about the issue of data tampering during the calculation process, for example, when using a CSV file as input in Id3gini. My idea is to save the hash value of the data using blockchain before performing calculations, and then check whether the hash value of the data has changed after the calculations to ensure the security of the computation. Is there any function in the mpyc development package that can be used to detect the hash value of the computed data? |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
I'm note sure about the scenario here. In the demo df = df[mpc.pid::m] # simple partition of dataset between m parties To validate these private inputs against changes during the computation does not seem very relevant? |
Beta Was this translation helpful? Give feedback.
I'm note sure about the scenario here. In the demo
id3gini.py
all parties start out with a copy of the same dataset stored in a.csv
file. But that's only to keep the demo simple. A realistic scenario is when all parties use a private.csv
file and secret-share these between each other usingmpc.input()
to create a horizontally partitioned dataset. See thekmsurvival.py
demo for an example, which is still a demo as each party uses a fragment of a globally available dataset, like thisTo validate these private inputs against changes during the computation does not seem very relevant?