Evaluating and Improving Child-Directed Automatic Speech Recognition

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 615–619 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 615 “Alexa in the wild” – Collecting unconstrained conversations with a modern voice assistant in a public environment Ingo Siegert Mobile Dialog Systems, Otto von Guericke University Magdeburg 39106 Magdeburg, Germany ingo.siegert@ovgu.de Abstract Datasets featuring modern voice assistants such as Alexa, Siri, Cortana and others allow an easy study of human-machine interactions. But data collections offering an unconstrained, unscripted public interaction are quite rare. Many studies so far have focused on private usage, short pre-defined task or specific domains. This contribution presents a dataset providing a large amount of unconstrained public interactions with a voice assistant. Up to now around 40 hours of device directed utterances were collected during a science exhibition touring through Germany. The data recording was part of an exhibit that engages visitors to interact with a commercial voice assistant system (Amazon’s ALEXA), but did not restrict them to a specific topic. A specifically developed quiz was starting point of the conversation, as the voice assistant was presented to the visitors as a possible joker for the quiz. But the visitors were not forced to solve the quiz with the help of the voice assistant and thus many visitors had an open conversation. The provided dataset – Voice Assistant Conversations in the wild (VACW) – includes the transcripts of both visitors requests and Alexa answers, identified topics and sessions as well as acoustic characteristics automatically extractable from the visitors’ audio files. Keywords: Intelligent Personal Assistants, Digital Personal Assistants, Voice Assistants, Conversational Agent, Smart Speaker, Amazon Echo, Amazon Alexa, public recordings, unconstrained interactions 1. Introduction In recent years, the market for commercial voice assistants has rapidly grown: e.g. Microsoft Cortana had 133 million active users in 2016 (Osborne, 2016) and the Echo Dot was the best-selling product on all of Amazon’s products in the last three holiday seasons (Kinsella, 2018). Furthermore, 72% of people who own a voice-activated speaker say their devices are often used as part of their daily routine (Klein- berg, 2018). Already in 2018 approximately 10% of the internet population used voice control according to (Jeffs, 20178). Mostly the ease of use is responsible for the attrac- tiveness of today’s voice assistant systems. Using nothing but speech commands, users can play music, search the web, create to-do lists, shop online, get instant weather forecasts, and control popular smart-home products. In future, the demands and possibilities will increase even more. Voice as- sistants have become one of the mainstays, with well-known examples like Alexa, Siri, or Cortana from Amazon, Apple, or Microsoft respectively (Osborne, 2016; Kinsella, 2018; Kleinberg, 2018). One can see the prevalence in different areas, such as Smart Home Control, Mobile Assistance, or Operating Systems. One of the main reasons behind this is the given naturalness of speaking as a form of commu- nication in contrast to additional periphery, and on top of that, the independence from …

Content truncated to 500 words