Abstract: In the recent past, modular robot assembly and metamorphosis has been evolved using gene regulatory networks. However, until now, no methodology exists to engineer such a regulatory network. Three existing representations will be employed to describe robot metamorphosis. A graph rewriting grammar describes state and connectivity transitions between robot organisms at the most abstract level. A communicating ï¬nite state machine introduces messages at an intermediate level. A regulatory network presents the process of metamorphosis at its least abstract level. In short, we present a design methodology for metamorphosis for which, as yet, only evolutionary methods did exist.
Abstract: In the ï¬eld of reservoir computing echo state networks (ESNs) and liquid state machines (LSMs) are the most commonly used networks. Comparative studies on these reservoirs identify the LSM as the network that yields the highest performance for speech recognition. But LSMs are not always usable in a real-time setting due to the computational costs of a large reservoir with spiking neurons.
In this paper a vowel classiï¬cation system is presented which consists of an ESN which processes cochlear ï¬ltered audio. The performance of the system is tested on a vowel classiï¬cation task using different signal-to-noise ratios (SNRs). The usefulness of this method is measured by comparing it to formant based vowel classiï¬cation systems. Results show that this ESN based system can get a performance similar to formant based vowel classiï¬cation systems on the clean dataset with only a small reservoir and even outperforms these methods on the noisy dataset
