Friday, March 29, 2024

Gray Wolf Optimizer — How It Can Be Used with Laptop Imaginative and prescient | by James Koh, PhD | Feb, 2024

Must read


As a bonus, get the code to use characteristic extraction anyplace

Towards Data Science
Picture created by DALL·E 3 primarily based on the immediate “Draw a pack of futuristic gray wolves at evening by the seaside.”

That is the final a part of my sequence of nature-inspired articles. Earlier, I had talked about algorithms impressed by genetics, swarm, bees, and ants. Right this moment, I’ll discuss wolves.

When a journal paper has a quotation rely spanning 5 figures, there’s some critical enterprise happening. Gray Wolf Optimizer [1] (GWO) is one such instance.

Like Particle Swarm Optimization (PSO), Synthetic Bee Colony (ABC), and Ant Colony Optimization (ACO), GWO can be a meta-heuristic. Though there’s no mathematical ensures to the answer, it really works effectively in apply and doesn’t require any analytical data of the underlying downside. This enables us to question from a ‘blackbox’, and easily make use to the noticed outcomes to refine our answer.

As talked about in my ACO article, all these finally relate again to the elemental idea of explore-exploit trade-off. Why, then, are there so many various meta-heuristics?

Firstly, it’s as a result of researchers need to publish papers. A great a part of their job entails exploring issues from totally different angles and sharing the methods during which their findings result in advantages over current approaches. (Or as some would say, publishing papers to justify their salaries and search promotions. However let’s not get there.)

Secondly, it’s because of the ‘No Free Lunch’ theorem [2] which the authors of GWO themselves talked about. Whereas that theorem was particularly saying there’s no free lunch for optimization algorithms, I believe it’s honest to say that the identical is true for Information Science normally. There isn’t any single final one-size-fits-all answer, and we regularly need to strive totally different approaches to see what works.

Due to this fact, let’s proceed so as to add yet one more meta-heuristic to our toolbox. As a result of it by no means hurts to have one other device which could come in useful in the future.

First, let’s take into account a easy classification downside on pictures. A intelligent method is to make use of pre-trained deep neural networks as characteristic extractors, to transform…



Supply hyperlink

More articles

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Latest article