Genetic algorithms are a form of machine learning that is focused on optimizing a particular output or outputs based on successive production of derived equations. The approach can be useful for optimizing a particular result when no training data are available, and when the optimization isn't known mathematically. These algorithms use the combination of selection, recombination, and mutation to evolve a solution to a problem. At a descriptive level, the intent of a genetic algorithm is to make a set of attempts at solving a problem using randomly selected equations. The first time through, it's likely that none of the solutions are very good.

What is an evolutionary algorithm?

AI is moving at a stellar speed and is probably one of most complex and present sciences. The complexity here is not meant as a level of difficulty in understanding and innovating (although of course, this is quite high), but as the degree of interrelation with other fields apparently disconnected. There are basically two schools of thought on how an AI should be properly built: the Connectionists start from the assumption that we should draw inspiration from the neural networks of the human brain, while the Symbolists prefer to move from banks of knowledge and fixed rules on how the world works. Given these two pillars, they think it is possible to build a system capable of reasoning and interpreting. In addition, a strong dichotomy is naturally taking shape in terms of problem-solving strategy: you can solve a problem through a simpler algorithm, which though it increases its accuracy in time (iteration approach), or you can divide the problem into smaller and smaller blocks (parallel sequential decomposition approach).

I suggested earlier that a third degree polynomial would be more difficult to generate automatically. In fact, what I found is that I would be lucky if the program converged at all. As time passed and the number of runs increased beyond a 100 then things started to slow down. Further, the fitness decrements became smaller and more infrequent. The give away for me was the runs slowing down.

Loyal readers will recall that last spring we conspired with artificial intelligence expert Randal Olson to develop the ultimate U.S. road trip. The map Olson came up with -- he did all the work, really -- optimized the best way to drive by car to 50 major U.S. landmarks, using machine learning algorithms and Google Maps. We're happy to report that Olson is back at it, just in time for summer road tripping. By leveraging the power of genetic algorithms and other artificial intelligence technology, Olsen's optimized loop route will get you across the country and back in a little over eight days -- starting in Concord, N.H., and dropping you back in Boston, Mass. How did Olson generate his road trip map?

If you read this blog regularly, you probably don't need an introduction to CRO or A/B testing. You know the major players, best practices, and you've likely tested your fair share of ideas. But, as an expert, you likely know some of the persistent frustrations with current approaches. In fact, only one in seven tests are actually successful. But new advances in artificial intelligence might help.

Genetic algorithms (GA) are a family of heuristics which are empirically good at providing a decent answer in many cases, although they are rarely the best option for a given domain. You mention derivative-based algorithms, but even in the absence of derivatives there are plenty of derivative-free optimization algorithms that perform way better than GAs. See this and this answer for some ideas. What many standard optimization algorithms have in common (even derivative-free methods) is the assumption that the underlying space is a smooth manifold (perhaps with a few discrete dimensions), and the function to optimize is somewhat well-behaved. However, not all functions are defined on a smooth manifold.

We explore the relation between memcomputing, namely computing with and in memory, and swarm intelligence algorithms. In particular, we show that one can design memristive networks to solve short-path optimization problems that can also be solved by ant-colony algorithms. By employing appropriate memristive elements one can demonstrate an almost one-to-one correspondence between memcomputing and ant colony optimization approaches. However, the memristive network has the capability of finding the solution in one deterministic step, compared to the stochastic multi-step ant colony optimization. This result paves the way for nanoscale hardware implementations of several swarm intelligence algorithms that are presently explored, from scheduling problems to robotics.

Water saturation is an important property in reservoir engineering domain. Thus, satisfactory classification of water saturation from seismic attributes is beneficial for reservoir characterization. However, diverse and non-linear nature of subsurface attributes makes the classification task difficult. In this context, this paper proposes a generalized Support Vector Data Description (SVDD) based novel classification framework to classify water saturation into two classes (Class high and Class low) from three seismic attributes seismic impedance, amplitude envelop, and seismic sweetness. G-metric means and program execution time are used to quantify the performance of the proposed framework along with established supervised classifiers. The documented results imply that the proposed framework is superior to existing classifiers. The present study is envisioned to contribute in further reservoir modeling.

It is a massive trial-and-error process. From time to time, you will hear about a new fantastic piece of computer science. For example, right now deep learning is the hot new thing. Some years ago, people were very excited about MapReduce. As an ecosystem changes, some tools become less likely to be useful while others gain dominance in common use cases.