Sorry, your browser cannot access this site
This page requires browser support (enable) JavaScript
Learn more >

LUYU LIU'S BLOG

Use Tucker-2 Decomposition on convolutional layer to compress CNNs. We test this compression techniques on two data sets: FashionMNIST and CIFAR-10. Check it out here Github. Convolutional Kernel T...

Authors

  • Zishu Liu
  • Yuxin Xie
  • Jichuan Zhang

This is an outline of emerging technology in distributed ledgers. If you want to know more about one topic, please refer to associated article. Lecture 1 Cryptographic Cryptographic Hash Functions...

Introduction Gabled circuits express the function to be computed as an “encrypted” boolean circuit, e.g. with ANDs, ORs, XORs, etc Garbled circuits are more efficient than for 2-party joint proces...

MPC (Multi-party Computation) Definition Given \(n\) parties, design a protocol which computes a public function \(f\) over inputs from each party such that the inputs remain private unless they wo...

Definition Neighboring dataset For any dataset \(D,D^{\prime}\in\mathcal{D}\), \(D,D^{\prime}\) are neighboring datasets if and only if one of the following two statements is true: \(\exists x\in ...

Introduction Query-based systems aim at giving researchers and organizations anonymous access to the data without sharing with them the individual-level data This can be through online interfaces,...

Pseudonymization To remove all direct identifiers, e.g. name, phone number, address, social security number etc, and just give each person a unique id that cannot be linked to them: Lookup table ...

许久不见啦!去年十月开始我忙于申请,然后紧接着又是毕业的相关事宜,以致我许久没有更新。我正在着手写一点关于我在清华四年的东西,希望能尽快发上来。与此同时,我也可能发一些我在IC的笔记(和以前一样)。 另外,我更新了博客的主题,希望大家都能喜欢 :)
Life

Multi Armed Bandit

  • \(t=1,2,\cdots,N\)
  • \(k\) arm
  • at \(t\), choose arm \(i\), receive reward \(X_{i,t}\sim{}D_i\)(iid unknown distribution)
  • Policy: \(\pi=\{i(t),t=1,2,\cdots,N\}\)
  • Goal: \(\max\mathbb{E}\{\sum_{t=1}^NX_{i,t}\}\)
  • \({\rm OPT}=N\max_i\mu_i=N\mu,\mu_i=\mathbb{E}\{X_{i,t}\}\)
  • \({\rm Reg}={\rm OPT}-\mathbb{E}\{\sum_{t=1}^NX_{i,t}|\pi\}\)
Math