In this article, we present the elitist particle filter based on evolutionary strategies (EPFES) as an efficient approach for nonlinear system identification. The EPFES is derived from the frequently-employed state-space model, where the relevant information of the nonlinear system is captured by an unknown state vector. Similar to classical particle filtering, the EPFES consists of a set of particles and respective weights which represent different realizations of the latent state vector and their likelihood of being the solution of the optimization problem. As main innovation, the EPFES includes an evolutionary elitist-particle selection which combines long-term information with instantaneous sampling from an approximated continuous posterior distribution. In this article, we propose two advancements of the previously-published elitist-particle selection process. Further, the EPFES is shown to be a generalization of the widely-used Gaussian particle filter and thus evaluated with respect to the latter for two completely different scenarios: First, we consider the so-called univariate nonstationary growth model with time-variant latent state variable, where the evolutionary selection of elitist particles is evaluated for non-recursively calculated particle weights. Second, the problem of nonlinear acoustic echo cancellation is addressed in a simulated scenario with speech as input signal: By using long-term fitness measures, we highlight the efficacy of the well-generalizing EPFES in estimating the nonlinear system even for large search spaces. Finally, we illustrate similarities between the EPFES and evolutionary algorithms to outline future improvements by fusing the achievements of both fields of research.

Horse betting is harder than it looks. At the 142nd Kentucky Derby last week, only one of five experts from Churchill Downs Racetrack correctly predicted the winner. None of them correctly predicted the top four horses. Known as a superfecta, this latter bet came with 540 to 1 odds, meaning 100 down would return 540,000. And although the experts failed to predict the finishing order, an anonymous group of internet users did.

Automatic modulation classification (AMC) is an important task for modern communication systems; however, it is a challenging problem when signal features and precise models for generating each modulation may be unknown. We present a new biologically-inspired AMC method without the need for models or manually specified features --- thus removing the requirement for expert prior knowledge. We accomplish this task using regularized stacked sparse denoising autoencoders (SSDAs). Our method selects efficient classification features directly from raw in-phase/quadrature (I/Q) radio signals in an unsupervised manner. These features are then used to construct higher-complexity abstract features which can be used for automatic modulation classification. We demonstrate this process using a dataset generated with a software defined radio, consisting of random input bits encoded in 100-sample segments of various common digital radio modulations. Our results show correct classification rates of > 99% at 7.5 dB signal-to-noise ratio (SNR) and > 92% at 0 dB SNR in a 6-way classification test. Our experiments demonstrate a dramatically new and broadly applicable mechanism for performing AMC and related tasks without the need for expert-defined or modulation-specific signal information.

UNU allows groups to chat online in a new way by forming a Swarm Intelligence that can answer questions, make predictions, and each decisions. Artificial intelligence (unanimous) UNU has not only conquered the Oscars and Super Bowl, but also the famous Kentucky derby. Its Swarm Intelligence software has reportedly made betting gamblers richer, with correct predictions owing to its successful forecasting methodologies. According to News Discovery, UNU is a software program that harnesses the collective power of horse racing professionals to correctly predict the first four horses that cross the finish line and in which order. At last weekend's Kentucky Derby, one experimental player witnessed the power of this software at first hand.

If you're going down to the racetrack, you might want to have an AI by your side. An artificial intelligence program developed by Unanimous A.I. successfully predicted the Superfecta at the 142nd Kentucky Derby last Saturday, turning a 20 bet into nearly 11,000. Using'Swarm Intelligence,' the AI was able to correctly choose the winning horse, Nyquist – along with the second, third, and fourth finishers. Unanimous AI's platform works by using the knowledge a group of people online, all logged into an the same interface where they can answer a series of questions. This animation shows how UNU's swarm intelligence makes its predictions Unanimous AI's platform works by using the knowledge a group of people online, all logged into an the same interface where they can answer a series of questions.

Picking the winner from the 20 horse field at the Kentucky Derby is hard. So hard, in fact, that no expert polled by SBNation was able to do it. That's why the holy grail at the racetrack is the Superfecta, where bettors are asked not only to pick the winner, but the second, third and fourth horses to finish the Derby. This is fiendishly difficult task that, not surprisingly, defeated every expert at Churchill Downs, where no one predicted the top four horses correctly, much less in the correct order. In the world of AI, even Bing Predicts blew it, picking only heavily favored Nyquist to win the race, but missing the other 3 picks entirely.

We present an approach to creating conversational agents that are capable of returning appropriate responses to natural language input. The approach described consists of a genetic algorithm used as a feature selection technique to evolve a subset of random features towards a set of features that are more relevant to the language used in the domain; therefore improving the conversational agent's ability to return appropriate responses. The results show that over multiple iterations of the evolutionary process the genetic algorithm was able to filter out unfit features. After the evolutionary process the features that were found to be relevant were tested on an unseen test set and the algorithm achieved an accuracy of 72.678%

Planning air missions for a team flying in hostile environments is a complex task, since multiple interrelated goals need to be considered, e.g., performing the mission tasks and avoiding enemy fire. The target assignment and route planning for the team should therefore be performed in a coordinated way. The mission planner suggested in this work combines genetic algorithms and particle swarm optimization in order to solve these two problems in an interconnected manner. Simulations are used for testing and analyzing the approach. It is concluded that the mission planner is able to suggest suitable plans in complex scenarios with three interrelated objectives: low risk exposure, high mission effectiveness and short route length.

For those betting on the 142nd Kentucky Derby on Saturday, there are several ways to approach the strategy. Last year, Jimmy Fallon's puppies took a stab at it--and correctly predicted the winner, American Pharoah. Or, you could rely on the experts from the Bleacher Report. Maybe you want to study up on your own, or see which horses are looking good that day. Go with TechRepublic's Steve Ranger on an inside look at the gold-plated gadget market that's received a big boost from Apple.

Face recall is a basic human cognitive process performed routinely, e.g., when meeting someone and determining if we have met that person before. Assisting a subject during face recall by suggesting candidate faces can be challenging. One of the reasons is that the search space - the face space - is quite large and lacks structure. A commercial application of face recall is facial composite systems - such as Identikit, PhotoFIT, and CD-FIT - where a witness searches for an image of a face that resembles his memory of a particular offender. The inherent uncertainty and cost in the evaluation of the objective function, the large size and lack of structure of the search space, and the unavailability of the gradient concept makes this problem inappropriate for traditional optimization methods. In this paper we propose a novel evolutionary approach for searching the face space that can be used as a facial composite system. The approach is inspired by methods of Bayesian optimization and differs from other applications in the use of the skew-normal distribution as its acquisition function. This choice of acquisition function provides greater granularity, with regularized, conservative, and realistic results.