Alex Cristia (École normale supérieure) will present in the next lab meeting of Inbal Arnon's Language Learning and Processing Lab, next Monday (23.11) at 10:30

zoom link:
https://zoom.us/j/5511229730 

Title:

Small babies, big data: Partnering with citizen scientists to study early vocal development across languages and cultures

Abstract:

Much research has examined vocal development in children from WEIRD (Henrich et al., 2010) cultures learning mainly IndoEuropean languages (Slobin, 2014). This line of research suggests strong similarities in vocal development across diverse socio-economic status and language backgrounds. However, each study relies on transcriptions of the speech by a small number of infants,  causing us to lack the power to detect potential language development trajectories, and thus likely underestimate diversity. In a series of studies, we've started to explore how we can bring big data approaches to infant research. We use daylong recordings to capture children's input and output. Recordings are analyzed using automatized speech processing techniques, to base our description of individual children on hours, rather than minutes, of audio. In our most recent efforts, we build on the growing citizen science movement to improve those automated classifications. So far, we have found strong differences across children who were diagnosed with Angelman syndrome compared to low risk controls, and broadly similar patterns of development across multiple datasets, with some potential cross-linguistic differences