Realizing being imitated: Vowel mapping with clearer articulation

Abstract

Abstract-The previous approach to vowel imitation learning between a caregiver and an infant (robot) [1] has assumed that the robot can segment the caregiver's utterance into its phoneme category, where the caregiver always imitates the robot utterance. However, in real situations, the caregiver does not always imitate the robot utterance, nor the robot does have the phoneme category (no segmentation capability). This paper presents a method to solve these issues, a weakly-supervised learning along with auto-regulation, that is active selection of action and data with underdeveloped classifier. To cope with notalways imitation problem, a weakly-supervised learning method is applied that is capable to handle incompletely segmented samples (not perfectly imitated voices). Further, the regulation classifier of the imitated voices is recursively applied in order to select good vocal primitives and to segment caregiver's imitated voices that improve the performance of the classifier itself. The simulation results are shown and the future issues are given.

Empirical approaches for investigating the origins of structure in speech

“…Several computational models have modelled infant speech acquisition in interaction with caregivers in order to investigate the learning mechanisms behind this complex process. These methods mainly rely on associative learning where the infant associates its babbles into vocal responses by its caregiver (Miura, Katsushi, Yuichiro, & Minoru, 2008;Miura, Yoshikawa, & Asada, 2007;Howard & Messum, 2011; see the introduction of Rasilo and Räsänen, 2017, for a review of related studies). In the latest study by Rasilo & Räsänen (2017), a virtual infant learns to imitate the eight Finnish vowel sounds by learning to associating its (inaccurate) babbles (with an infant-sized vocal tract) with responses from adult human participants.…”

Biomimetic learning, not learning biomimetics: A survey of developmental learning

“…Completely unsupervised learning is generally not desired since results should conform to a desired language. For this reason, tutors or exemplars are often used to direct the learning [18]- [20]. One study uses children to teach virtual agents to accentuate co-development [21].…”