Learning to imitate adult speech with the KLAIR virtual infant

Pre-linguistic infants need to learn how to produce spoken word forms that have the appropriate intentional effect on adult carers. One proposed imitation strategy is based on the idea that infants are innately able to match the sounds of their own babble to sounds of adults, while another propos strategy requires only reinforcement signals from improve random imitations. Here we demonst knowledge gained from interactions between infants and adults can provide useful normalizing data that improves the recognisability of infant imitations. We use the KLAIR virtual infant toolkit to collect spoken interactions with adults, the collected data to learn adult-to-infant mappings, and construct imitations of adult utterances using KLAIR's articulatory synthesizer. We show that speakers reformulate KLAIR's productions in terms of standard phonological forms, and that these reformulations can be used to train a system that generates infant imitations recognisable to adults than a system based on babbling alone.