SLT2021 hosts the following challenges. The details are as follows:
Organizers: Youjun Xiong, Lei Xie, Huan Tan, Dongyan Huang, Jean-Marc Odobez, Petr Motlicek, Weipeng He, Yuexian Zou, Hui Bu, Jian Wu
Robots, as useful assistants and playmates, are becoming more and more popular in people's daily life. As the first chain of HRSI (Human Robot Speech Interaction), keyword spotting (KWS) technology (a.k.a. wake-up word detection), directly determines the experience of subsequent interactions. Meanwhile, the accuracy of sound source location (SSL) can provide essential cues for subsequent beamforming, speech enhancement and speech recognition algorithms. In home environments, the following interferences pose great challenges to HRSI: 1) various types of noises from TV, radio, other electrical appliances and human talking, 2) echoes from the loudspeaker equipped on the robot, 3) room reverberation and 4) noises from the mechanical movements of the robot (mechanical noise in short). These noise interferences complicate KWS and SSL to a great extent. Thus, robust algorithms are highly in demand.
Alpha-mini is an excellent robot produced by UBTECH, equipped with intelligent speech interaction module based on a 4-microphone array. As a flagship satellite event of the 2021 IEEE Spoken Language Technology (SLT) Workshop, Alpha-mini Speech Challenge (ASC) will provide a common benchmark for KWS, SSL and related tasks. We will also open source a sizable dataset for KWS and SSL based on Alpha-mini and a common platform, covering abundant indoor noise, echo and reverberation, to better tackle the problem of HSRI in real application scenarios. More details for this challenge can be found in asc.ubtrobot.com.
Organizers: Lei Xie, Zhijian Ou, Qingyang Hong, Xiulin Li, Bo Liu
With the rapid development of automatic speech recognition technology and its widespread use, children speech recognition has an exponential range of applications. It plays a more and more important role in applications, such as intelligent education, toys, smart speakers, etc. However, the speech and language characteristics of children's voice are substantially different from those of adults, which poses challenges to the current ASR technology. Another bottleneck to the current development of children speech recognition is the lack of labelled children's speech in the research community.
We launch the Children Speech Recognition Challenge (CSRC), as a flagship satellite event of IEEE SLT 2021 workshop. The Challenge will release about 400 hours of data for registered teams and set up two challenge tracks, targeting to boost children speech recognition research. Hopefully, the Challenge will also provide a good testbed for related techniques such as domain adaptation, transfer learning, and so on.
The challenge organizer will provide ~340h adult reading data, ~28.6h children reading data and ~29.5h children dialogue data as well as their corresponding transcripts. Two tracks are open for participants, and the difference between the two tracks is that only the officially provided data are allowed for model training in track 1, while track 2 allows participants to use external data listed in openslr in addition to the officially provided data for system building. More details for this challenge can be found from https://www.data-baker.com/csrc_challenge.html
Copyright © 2019-2021. SLT2021 Organizing Committee. All rights reserved.