e success of deep learning depends on nding an architecture to t the task. As deep learning has scaled up to more challenging tasks, the architectures have become di cult to design by hand.is 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.
We consider the Dynamic Map Visitation Problem (DMVP), in which a team of agents must visit a collection of critical locations as quickly as possible, in an environment that may change rapidly and unpredictably during the agents' navigation. We apply recent formulations of time-varying graphs (TVGs) to DMVP, shedding new light on the computational hierarchy R ⊃ B ⊃ P of TVG classes by analyzing them in the context of graph navigation. We provide hardness results for all three classes, and for several restricted topologies, we show a separation between the classes by showing severe inapproximability in R, limited approximability in B, and tractability in P. We also give topologies in which DMVP in R is fixed parameter tractable, which may serve as a first step toward fully characterizing the features that make DMVP
No abstract
Multitask learning, i.e. learning several tasks at once with the same neural network, can improve performance in each of the tasks. Designing deep neural network architectures for multitask learning is a challenge: There are many ways to tie the tasks together, and the design choices matter. The size and complexity of this problem exceeds human design ability, making it a compelling domain for evolutionary optimization. Using the existing state of the art soft ordering architecture as the starting point, methods for evolving the modules of this architecture and for evolving the overall topology or routing between modules are evaluated in this paper. A synergetic approach of evolving custom routings with evolved, shared modules for each task is found to be very powerful, significantly improving the state of the art in the Omniglot multitask, multialphabet character recognition domain. This result demonstrates how evolution can be instrumental in advancing deep neural network and complex system design in general.MTL [3] exploits relationships across problems to increase overall performance. The underlying idea is that if multiple tasks are related, the optimal models for those tasks will be related as well. In the convex optimization setting, this idea has been implemented via various regularization penalties on shared parameter matrices [1,7,18,22]. Evolutionary methods have also had success in MTL, especially in sequential decision-making domains [13,16,19,38,41].
Novelty search and related diversity-driven algorithms provide a promising approach to overcoming deception in complex domains. The behavior characterization (BC) is a critical choice in the application of such algorithms. The BC maps each evaluated individual to a behavior, i.e., some vector representation of what the individual is or does during evaluation. Search is then driven towards diversity in a metric space of these behaviors. BCs are built from handdesigned features that are limited by human expertise, or upon generic descriptors that cannot exploit domain nuance. The main contribution of this paper is an approach that addresses these shortcomings. Generic behaviors are recorded from evolution on several training tasks, and a new BC is learned from them that funnels evolution towards successful behaviors on any further tasks drawn from the domain. This approach is tested in increasingly complex simulated mazesolving domains, where it outperforms both hand-coded and generic BCs, in addition to outperforming objective-based search. The conclusion is that adaptive BCs can improve search in many-task domains with little human expertise.
Several models have been developed to predict how the COVID-19 pandemic spreads, and how it could be contained with nonpharmaceutical interventions, such as social distancing restrictions and school and business closures. This article demonstrates how evolutionary AI can be used to facilitate the next step, i.e., determining most effective intervention strategies automatically. Through evolutionary surrogate-assisted prescription, it is possible to generate a large number of candidate strategies and evaluate them with predictive models. In principle, strategies can be customized for different countries and locales, and balance the need to contain the pandemic and the need to minimize their economic impact. Early experiments suggest that workplace and school restrictions are the most important and need to be designed carefully. They also demonstrate that results of lifting restrictions can be unreliable, and suggest creative ways in which restrictions can be implemented softly, e.g., by alternating them over time. As more data becomes available, the approach can be increasingly useful in dealing with COVID-19 as well as possible future pandemics.
Deep neural networks (DNNs) have produced state-of-the-art results in many benchmarks and problem domains. However, the success of DNNs depends on the proper configuration of its architecture and hyperparameters. Such a configuration is difficult and as a result, DNNs are often not used to their full potential. In addition, DNNs in commercial applications often need to satisfy real-world design constraints such as size or number of parameters. To make configuration easier, automatic machine learning (AutoML) systems for deep learning have been developed, focusing mostly on optimization of hyperparameters.This paper takes AutoML a step further. It introduces an evolutionary AutoML framework called LEAF that not only optimizes hyperparameters but also network architectures and the size of the network. LEAF makes use of both state-of-the-art evolutionary algorithms (EAs) and distributed computing frameworks. Experimental results on medical image classification and natural language analysis show that the framework can be used to achieve state-of-the-art performance. In particular, LEAF demonstrates that architecture optimization provides a significant boost over hyperparameter optimization, and that networks can be minimized at the same time with little drop in performance. LEAF therefore forms a foundation for democratizing and improving AI, as well as making AI practical in future applications.
Food production in conventional agriculture faces numerous challenges such as reducing waste, meeting demand, maintaining flavor, and providing nutrition. Contained environments under artificial climate control, or cyber-agriculture, could in principle be used to meet many of these challenges. Through such environments, phenotypic expression of the plant—mass, edible yield, flavor, and nutrients—can be actuated through a “climate recipe,” where light, water, nutrients, temperature, and other climate and ecological variables are optimized to achieve a desired result. This paper describes a method for doing this optimization for the desired result of flavor by combining cyber-agriculture, metabolomic phenotype (chemotype) measurements, and machine learning. In a pilot experiment, (1) environmental conditions, i.e. photoperiod and ultraviolet (UV) light (known to affect production of flavor-active molecules in edible plants) were applied under different regimes to basil plants ( Ocimum basilicum ) growing inside a hydroponic farm with an open-source design; (2) flavor-active volatile molecules were measured in each plant using gas chromatography-mass spectrometry (GC-MS); and (3) symbolic regression was used to construct a surrogate model of this chemistry from the input environmental variables, and this model was used to discover new combinations of photoperiod and UV light to increase this chemistry. These new combinations, or climate recipes, were then implemented in the hydroponic farm, and several of them resulted in a marked increase in volatiles over control. The process also led to two important insights: it demonstrated a “dilution effect”, i.e. a negative correlation between weight and desirable chemical species, and it discovered the surprising effect that a 24-hour photoperiod of photosynthetic-active radiation, the equivalent of all-day light, induces the most flavor molecule production in basil. In this manner, surrogate optimization through machine learning can be used to discover effective recipes for cyber-agriculture that would be difficult and time-consuming to find using hand-designed experiments.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.