How could this possibly be faster than a linear classifier? In "Implementation" near Section 2 they claim that their implementation is better than a linear classifier but that seems like it couldn't possibly be true could it? A single floating point operation per parameter has gotta be faster than multiple branches, right?
A linear classifier can only handle linearly separable things, which limits prediction accuracy a lot.
I don't think you'll get anywhere close to this level of classification with a linear model. In the experiments it is shown to outperform SVM RBF.
The platforms in question doesn't have hardware for floating point, any floating point support needs to be emulated in software (slow).
Sadly it was a while ago - the closest I could find was a MIT one mentioning essay graders judging by length and connective words allowing for properly structured nonsense to be judged as good writing.
Are you referring to animals (including, under the right conditions, humans) seeing the polarization of light?
That allows them to figure out the position of the Sun when it is not visible due to clouds, which is indeed useful for navigating.
But what this article is about is seeing the Earth's magnetic field.
At least some migratory birds are able to detect both their longitude and latitude [1] [2], which requires more than just knowing where the Sun is.
Humans have been able to figure out latitude for a very long time...at least as far back as the ancient Greeks, and probably much farther back.
Longitude, on the other hand, eluded us until we were able to make reasonably accurate clocks. For sea navigation, that wasn't until the 18th century, long after we had compasses and knew about polarized light.
>...the inaccuracies stemmed primarily from a statistical and reporting/formatting error that led to further inaccuracies.
>The review found that there was no intended deception or evidence of deliberate misconduct, and that the significance of the results and discussion in the article would not change because of the errors.
>However, since the number of errors is too voluminous to be executed by issuing a correction statement, the journal is withdrawing the article and will republish it as a corrected version in a subsequent issue, and will utilize the same DOI as the originally published version of the article. The authors agree with this decision.
