English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Majority of Artificial Neural Network (ANN) implementations in autonomoussystems use a fixed/user-prescribed network topology, leading to sub-optimalperformance and low portability. The existing neuro-evolution of augmentingtopology or NEAT paradigm offers a powerful alternative by allowing the networktopology and the connection weights to be simultaneously optimized through anevolutionary process. However, most NEAT implementations allow theconsideration of only a single objective. There also persists the question ofhow to tractably introduce topological diversification that mitigatesoverfitting to training scenarios. To address these gaps, this paper develops amulti-objective neuro-evolution algorithm. While adopting the basic elements ofNEAT, important modifications are made to the selection, speciation, andmutation processes. With the backdrop of small-robot path-planningapplications, an experience-gain criterion is derived to encapsulate the amountof diverse local environment encountered by the system. This criterionfacilitates the evolution of genes that support exploration, thereby seeking togeneralize from a smaller set of mission scenarios than possible withperformance maximization alone. The effectiveness of the single-objective(optimizing performance) and the multi-objective (optimizing performance andexperience-gain) neuro-evolution approaches are evaluated on two differentsmall-robot cases, with ANNs obtained by the multi-objective optimizationobserved to provide superior performance in unseen scenarios.

Multi-criteria Evolution of Neural Network Topologies: Balancing Experience and Performance in Autonomous Systems