0

I'm reviewing an algorithm and it all makes sense except for this line:

$$ \boldsymbol{b} \leftarrow (||\boldsymbol{b}|| \_{2} - \beta)_{+} \widehat{\boldsymbol{b}} $$

$\boldsymbol b$ is an $N\times 1$ vector and $\beta$ is a constant. I'm not sure what the subscript plus denotes.

Henrik supports the community
  • 6,831
Arik D. Brown
  • 1
  • 1
    Please clarify your specific problem or provide additional details to highlight exactly what you need. As it's currently written, it's hard to tell exactly what you're asking. – Community Oct 20 '23 at 15:12
  • 1
    $(x)+$ usually denotes $\max{x, 0}$. Related topics are hinge loss and [ReLU](https://en.wikipedia.org/wiki/Rectifier(neural_networks)). – angryavian Oct 20 '23 at 15:25
  • You found a fault in the documentation of that algorithm. It used unexplained notation that was unknown to the reader (you). – GEdgar Oct 20 '23 at 16:43
  • The _2 is supposed to be a subscript 2. Not sure why it didn't translate when posted. The bracketed b vector with the _2 is supposed to be the norm of the vector that the constant beta is subtracted from. Since the norm gives scalar I'm not sure how this gets assigned to the vector b. The subscripted plus sign I'm not sure of either. This is the last step in the algorithm to give the iteratively optimized vector b. – Arik D. Brown Oct 20 '23 at 18:13
  • Please edit the question instead of posting corrections in a comment! – Henrik supports the community Oct 20 '23 at 22:30

0 Answers0