Why is Bootstrapping important?
Generally speaking, homomorphic encryption software offers a limited number of operations and with bootstrapping we’re able to exceed this limitation, meaning we can do an unlimited number of operations while the data is still in the encrypted state.
What separates your method from previous methods?
With our method we have changed the middle step in bootstrapping. There is a process called Homomorphic Bit Extraction. In our process, our method has lower depth and lower complexity. In our method, we use one more polynomial which is called Lowest Digit Removal Polynomial. In previous methods they used 3 kinds of function and our method uses 4 functions.
What is the advantage to your method?
It quicker and with smaller parameters
Are there any real life applications for your improved method?
In the cases of machine learning and data analysis in the encrypted states, you would need large depth computation of encrypted data. So we need to insert bootstrapping steps for every 5 or 6 situations so that the performance of bootstrapping goes well then this kind of application will go well. If someone wants to use homomorphic encryption in the future, then bootstrapping will be necessary. Maybe if you are just using a simple application you wouldn’t need any bootstrapping. But usually, when you have large depth computation the bootstrapping will be necessary. Here is the complete presentation that Kyoohyung Han created for his talk on Lower Digit Removal and Improved FHE Bootstrapping at EuroCrypt 2018
