In recent years, the automatic speech recognition technology has witnessed breakthrough developments. People are increasingly interacting with this technology on a daily basis when using smart speaker devices such as Alexa, Siri, etc. Despite this progress, a problem waiting to be solved is automatic detection and recognition of stuttered events in spoken conversations. Addressing this problem is critical in enabling the developments in speech recognition technology to reach individuals who have atypical speaking style. Recognizing the importance of this, Project Boli aims to address the challenge by a three-fold approach.

• Firstly, we are creating a dataset by reaching out to individuals who stutter.
• Secondly, this dataset will be analyzed to further the scientific understanding on the acoustic patterns in speech signals.
• Thirdly, this dataset will be used to further R&D on developing signal processing and machine algorithms to detect and classify atypical patterns in stuttered speech.

We are thankful for your assistance in contributing to creation of the dataset and/or spreading the word about this project to individuals who can contribute to it. This will help in the progress of science and technology for the benefit of humanity.