this is for holding javascript data
Richard Guo edited subsection_Non_convex_optimization_formulation__.tex
over 8 years ago
Commit id: fab354b381ec93bfa8e45006896f8c7bc3fa3ea2
deletions | additions
diff --git a/subsection_Non_convex_optimization_formulation__.tex b/subsection_Non_convex_optimization_formulation__.tex
index 1a534b0..7d846eb 100644
--- a/subsection_Non_convex_optimization_formulation__.tex
+++ b/subsection_Non_convex_optimization_formulation__.tex
...
However, this is discontinuous and the scale of $\mathbf{q}$ is unidentifiable. The problem above is recast as
\[ \min f(\mathbf{q}; \hat{\mathbf{Y}}) = \frac{1}{p} \sum_{k=1}^p h_{\mu} (\mathbf{q}^T \hat{\mathbf{y}_k}), \quad s.t. \ \|q\|_2 = 1, \]
where $\hat{\mathbf{Y}}$ is the data after pre-conditioning. The loss function is a special convex function,
namely
\[ h_{\mu} (z) = \mu \log \cosh(z /
\mu) \mu). \]
Next, we discuss some of the key properties of this problem.