[mlpack-git] [mlpack/mlpack] QUIC-SVD fails if m=n. (#717)

[Marcus Edel]

The QUIC-SVD method doesn't work if `m = n`, however as pointed out in the paper **QUIC-SVD: Fast SVD Using Cosine Trees**:

> Throughout this paper we assume m ≥ n, such that sampling rows gives bigger speedup than sampling columns.


it should work. One problem is that the CosineNodeSplit(); returns `if numColumns < 2` so the upcoming operations result in an error:

```
// Split the node into left and right children.
currentNode->CosineNodeSplit();

// Obtain pointers to the left and right children of the current node.
CosineTree *currentLeft, *currentRight;
currentLeft = currentNode->Left();
currentRight = currentNode->Right();
```

Right now, I can't see an easy fix, without changing some bigger parts of the code. Maybe someone else has a simple idea? Maybe @siddharth-agrawal  has an idea?

---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/mlpack/mlpack/issues/717
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160707/3ab2c080/attachment.html>