Weizmann Institute, The Mathematics of Modern Cryptography
this video https://www.youtube.com/watch?v=O8IvJAIvGJo
probably this event https://simons.berkeley.edu/talks/wichs-brakerski-2015-07-06
01:00 intro
02:55 homomorphic evaluation function
04:45 this is what it is supposed to provide
05:00 input privacy
06:28 compactness
07:00 Fully H.E.
08:00 universal set of gates, NAND
09:25 leveled fully homomorphic encryption, only bounded depth f
10:10 1978 to 2008 - from idea to 1st solution
11:10 approximate eigenvector method
13:22 insecure if exact eigenvector
13:55 tweak idea: approx eigenvector
15:35 test addition, works. but noise grows.
17:38 test multiplication, works only for C_2 small
20:13 LWE learning with errors, 2005 Regev
25:25 re-arrange notation
26:01 encryption scheme from from LWE
31:40 generalize from vectors to matrices
33:45 what do we need, small C_2
35:00 binary decomposition
39:00 whole scheme, invariant
42:00 decryption
44:00 homomorphic operations, e.g. multiplication
47:24 NAND, noise
50:10 homomorphic circuit evaluation
51:20 conclusion
Please fork, edit/correct/extend, and pull request. Thanks.