**Input and Initialization:**

Training data **X**, frame bound A,B, iteration *num*, gradient descent iterations *r*.

Build frame **<sup>Φ</sup>** <sup>∈</sup> <sup>R</sup>*N*×*M*and **<sup>Ψ</sup>** <sup>∈</sup> <sup>R</sup>*N*×*M*,either by using random entries, or using M randomly

chosen data.
