Known Uses

http://blog.otoro.net/2016/05/07/backprop-neat/ NEAT

http://eplex.cs.ucf.edu/papers/stanley_gpem07.pdf Compositional Pattern Producing Networks: A Novel Abstraction of Development

https://blog.evorithmics.org/2016/01/31/when-will-evolution-outperform-local-search/

http://arxiv.org/abs/1606.02580 Convolution by Evolution: Differentiable Pattern Producing Networks

Our main result is that DPPNs can be evolved/trained to compress the weights of a denoising autoencoder from 157684 to roughly 200 parameters, while achieving a reconstruction accuracy comparable to a fully connected network with more than two orders of magnitude more parameters.

http://eplex.cs.ucf.edu/papers/morse_gecco16.pdf Simple Evolutionary Optimization Can Rival Stochastic Gradient Descent in Neural Networks

This paper challenges these views, suggesting that EAs can be made to run significantly faster than previously thought by evaluating individuals only on a small number of training examples per generation. Surprisingly, using this approach with only a simple EA (called the limited evaluation EA or LEEA) is competitive with the performance of the state-of-the-art SGD variant RMSProp on several benchmarks with neural networks with over 1,000 weights.

As long as each step of evolution is not changing the behavior of each network in a radical way, this fitness inheritance builds up for those individuals who are more likely to be more globally fit than their peers

https://arxiv.org/pdf/1609.09106v1.pdf HYPERNETWORKS

https://arxiv.org/abs/1703.00548 Evolving Deep Neural Networks

This paper proposes an automated method, CoDeepNEAT, for optimizing deep learning architectures through evolution. By extending existing neuroevolution methods to topology, components, and hyperparameters, this method achieves results comparable to best human designs in standard benchmarks in object recognition and language modeling. It also supports building a real-world application of automated image captioning on a magazine website. Given the anticipated increases in available computing power, evolution of deep networks is promising approach to constructing deep learning applications in the future.

https://arxiv.org/pdf/1703.02806v1.pdf Deep Reservoir Computing Using Cellular Automata

Echo State Networks and Liquid State Machines have been proposed as possible RNN alternatives, under the name of Reservoir Computing (RC). RCs are far more easy to train. In this paper, Cellular Automata are used as reservoir, and are tested on the 5-bit memory task (a well known benchmark within the RC community). The work herein provides a method of mapping binary inputs from the task onto the automata, and a recurrent architecture for handling the sequential aspects of it. Furthermore, a layered (deep) reservoir architecture is proposed.

https://arxiv.org/abs/1703.01513v1 Genetic CNN

We run the genetic process on two small datasets, i.e., MNIST and CIFAR10, demonstrating its ability to evolve and find high-quality structures which are little studied before.

https://arxiv.org/abs/1703.03864 Evolution Strategies as a Scalable Alternative to Reinforcement Learning

We explore the use of Evolution Strategies, a class of black box optimization algorithms, as an alternative to popular RL techniques such as Q-learning and Policy Gradients. Experiments on MuJoCo and Atari show that ES is a viable solution strategy that scales extremely well with the number of CPUs available: By using hundreds to thousands of parallel workers, ES can solve 3D humanoid walking in 10 minutes and obtain competitive results on most Atari games after one hour of training time. In addition, we highlight several advantages of ES as a black box optimization technique: it is invariant to action frequency and delayed rewards, tolerant of extremely long horizons, and does not need temporal discounting or value function approximation.

http://www.evolvingai.org/PicbreederCanalization One hypothesized mechanism for evolvability is developmental canalization, wherein certain dimensions of variation become more likely to be traversed and others are prevented from being explored.

https://arxiv.org/abs/1704.05554 Discovering Evolutionary Stepping Stones through Behavior Domination Behavior domination is proposed as a tool for understanding and harnessing the power of evolutionary systems to discover and exploit useful stepping stones. Novelty search has shown promise in overcoming deception by collecting diverse stepping stones, and several algorithms have been proposed that combine novelty with a more traditional fitness measure to refocus search and help novelty search scale to more complex domains. However, combinations of novelty and fitness do not necessarily preserve the stepping stone discovery that novelty search affords. In several existing methods, competition between solutions can lead to an unintended loss of diversity. Behavior domination defines a class of algorithms that avoid this problem, while inheriting theoretical guarantees from multiobjective optimization. Several existing algorithms are shown to be in this class, and a new algorithm is introduced based on fast non-dominated sorting. Experimental results show that this algorithm outperforms existing approaches in domains that contain useful stepping stones, and its advantage is sustained with scale. The conclusion is that behavior domination can help illuminate the complex dynamics of behavior-driven search, and can thus lead to the design of more scalable and robust algorithms.

https://arxiv.org/abs/1706.04241 On Optimistic versus Randomized Exploration in Reinforcement Learning

Optimistic approaches that have been proposed in the literature sacrifice statistical efficiency for the sake of computational efficiency. Randomized approaches, on the other hand, may enable simultaneous statistical and computational efficiency.

http://www.evolvingai.org/files/how2017kouvaris.pdf How evolution learns to generalise: Using the principles of learning theory to understand the evolution of developmental organisation

https://arxiv.org/pdf/1703.01041.pdf Large-Scale Evolution of Image Classifiers

Despite significant computational requirements, we show that it is now possible to evolve models with accuracies within the range of those published in the last year.

https://arxiv.org/pdf/1710.04748.pdf HyperENTM: Evolving Scalable Neural Turing Machines through HyperNEAT

https://arxiv.org/pdf/1709.00268.pdf Algorithmically probable mutations reproduce aspects of evolution such as convergence rate, genetic memory, modularity, diversity explosions, and mass extinction

https://arxiv.org/abs/1712.06567 Deep Neuroevolution: Genetic Algorithms Are a Competitive Alternative for Training Deep Neural Networks for Reinforcement Learning

https://arxiv.org/abs/1712.06564v1 On the Relationship Between the OpenAI Evolution Strategy and Stochastic Gradient Descent

Incorporating these lessons, the paper concludes by demonstrating that ES can achieve 99% accuracy on MNIST, a number higher than any previously published result for any evolutionary method.

Even if ES falls short of a perfect gradient estimation, the ability to approximate a reasonable gradient may still be sufficient to compete in domains without perfect gradient information available with respect to final performance.

Directly perturbing policy parameters can offer the benefit of correlated exploration in action space.

A population of perturbations in effect contains more information than a single suggested gradient direction. Can such a population perhaps be leveraged to complement gradient knowledge? Can these kinds of computations be interleaved? Can they shed light on each other and perhaps yield a more fundamental understanding overall?

https://arxiv.org/abs/1709.05601v1 https://arxiv.org/abs/1709.05601v1

Markov Brains are a class of evolvable artificial neural networks (ANN). They differ from conventional ANNs in many aspects, but the key difference is that instead of a layered architecture, with each node performing the same function, Markov Brains are networks built from individual computational components. These computational components interact with each other, receive inputs from sensors, and control motor outputs. The function of the computational components, their connections to each other, as well as connections to sensors and motors are all subject to evolutionary optimization. Here we describe in detail how a Markov Brain works, what techniques can be used to study them, and how they can be evolved.

https://arxiv.org/abs/1802.01548 Regularized Evolution for Image Classifier Architecture Search

This study introduces a regularized version of a popular asynchronous evolutionary algorithm. We rigorously compare it to the non-regularized form and to a highly-successful reinforcement learning baseline. Using the same hardware, compute effort and neural network training code, we conduct repeated experiments side-by-side, exploring different datasets, search spaces and scales. We show regularized evolution consistently produces models with similar or higher accuracy, across a variety of contexts without need for re-tuning parameters. In addition, regularized evolution exhibits considerably better performance than reinforcement learning at early search stages, suggesting it may be the better choice when fewer compute resources are available. This constitutes the first controlled comparison of the two search algorithms in this context.

https://arxiv.org/abs/1803.00657 Evolutionary Generative Adversarial Networks

Unlike existing GANs, which employ a pre-defined adversarial objective function alternately training a generator and a discriminator, we utilize different adversarial training objectives as mutation operations and evolve a population of generators to adapt to the environment (i.e., the discriminator). We also utilize an evaluation mechanism to measure the quality and diversity of generated samples, such that only well-performing generator(s) are preserved and used for further training. In this way, E-GAN overcomes the limitations of an individual adversarial training objective and always preserves the best offspring, contributing to progress in and the success of GANs. Experiments on several datasets demonstrate that E-GAN achieves convincing generative performance and reduces the training problems inherent in existing GANs.

https://arxiv.org/abs/1803.03453 The Surprising Creativity of Digital Evolution: A Collection of Anecdotes from the Evolutionary Computation and Artificial Life Research Communities

This paper is the crowd-sourced product of researchers in the fields of artificial life and evolutionary computation who have provided first-hand accounts of such cases. It thus serves as a written, fact-checked collection of scientifically important and even entertaining stories. In doing so we also present here substantial evidence that the existence and importance of evolutionary surprises extends beyond the natural world, and may indeed be a universal property of all complex evolving systems.

https://arxiv.org/abs/1801.05159v1 GitGraph - Architecture Search Space Creation through Frequent Computational Subgraph Mining

Concretely, we (a) extract and publish GitGraph, a corpus of neural architectures and their descriptions; (b) we create problem-specific neural architecture search spaces, implemented as a textual search mechanism over GitGraph; © we propose a method of identifying unique common subgraphs within the architectures solving each problem (e.g., image processing, reinforcement learning), that can then serve as modules in the newly created problem specific neural search space.

https://eng.uber.com/deep-neuroevolution/

https://arxiv.org/pdf/1806.08099v1.pdf Lamarckian Evolution of Convolutional Neural Networks

Instead of randomly initializing the network weights before training, it is also possible to inherit the already learned weights of an ancestor network. This inheritance of aquired traits is a form of lamarckian evolution which, while rejected in biology, can prove useful in artificial evolution.

https://arxiv.org/abs/1806.05695v1 Evolving simple programs for playing Atari games

Cartesian Genetic Programming (CGP) has previously shown capabilities in image processing tasks by evolving programs with a function set specialized for computer vision. A similar approach can be applied to Atari playing. Programs are evolved using mixed type CGP with a function set suited for matrix operations, including image processing, but allowing for controller behavior to emerge. While the programs are relatively small, many controllers are competitive with state of the art methods for the Atari benchmark set and require less training time. By evaluating the programs of the best evolved individuals, simple but effective strategies can be found.