The INTERSPEECH 2017 Computational Paralinguistics Challenge addresses three different problems for the first time in research competition under well-defined conditions: In the Addressee sub-challenge, it has to be determined whether speech produced by an adult is directed towards another adult or towards a child; in the Cold sub-challenge, speech under cold has to be told apart from 'healthy' speech; and in the Snoring sub-challenge, four different types of snoring have to be classified. In this paper, we describe these sub-challenges, their conditions, and the baseline feature extraction and classifiers, which include data-learnt feature representations by end-to-end learning with convolutional and recurrent neural networks, and bag-of-audio-words for the first time in the challenge series.

Measurements of infants' quotidian experiences provide critical information about early development. However, the role of sampling methods in providing these measurements is rarely examined. Here we directly compare language input from hour-long video-recordings and daylong audio-recordings within the same group of 44 infants at 6 and 7 months. We compared 12 measures of language quantity and lexical diversity, talker variability, utterance-type, and object presence, finding moderate correlations across recording-types. However, video-recordings generally featured far denser noun input across these measures compared to the daylong audio-recordings, more akin to 'peak' audio hours (though not as high in talkers and word-types). Although audio-recordings captured ~10 times more awake-time than videos, the noun input in them was only 2-4 times greater. Notably, whether we compared videos to daylong audio-recordings or peak audio times, videos featured relatively fewer declaratives and more questions; furthermore, the most common video-recorded nouns were less consistent across families than the top audio-recording nouns were. Thus, hour-long videos and daylong audio-recordings revealed fairly divergent pictures of the language infants hear and learn from in their daily lives. We suggest that short video-recordings provide a dense and somewhat different sample of infants' language experiences, rather than a typical one, and should be used cautiously for extrapolation about common words, talkers, utterance-types, and contexts at larger timescales. If theories of language development are to be held accountable to 'facts on the ground' from observational data, greater care is needed to unpack the ramifications of sampling methods of early language input.

A range of demographic variables influence how much speech young children hear. However, because studies have used vastly different sampling methods, quantitative comparison of interlocking demographic effects has been nearly impossible, across or within studies. We harnessed a unique collection of existing naturalistic, day-long recordings from 61 homes across four North American cities to examine language input as a function of age, gender, and maternal education. We analyzed adult speech heard by 3- to 20-month-olds who wore audio recorders for an entire day. We annotated speaker gender and speech register (child-directed or adult-directed) for 10,861 utterances from female and male adults in these recordings. Examining age, gender, and maternal education collectively in this ecologically-valid dataset, we find several key results. First, the speaker gender imbalance in the input is striking: children heard 2–3× more speech from females than males. Second, children in higher-maternal-education homes heard more child-directed speech than those in lower-maternal education homes. Finally, our analyses revealed a previously unreported effect: the proportion of child-directed speech in the input increases with age, due to a decrease in adult-directed speech with age. This large-scale analysis is an important step forward in collectively examining demographic variables that influence early development, made possible by pooled, comparable, day-long recordings of children’s language environments. The audio recordings, annotations, and annotation software are readily available for re-use and re-analysis by other researchers.

share first authorship. Elika Bergelson, Andrei Amatuni, and Shannon Dailey were in Brain and Cognitive Sciences at the University of Rochester during data collection. Authors have no conflicts of interest to declare.