14,537 accept challenge
Mozilla Voice Challenge

Mozilla Voice Challenge

We are on a mission to find which specific open source technologies need to be developed and integrated to power voice-enabled products. Read Overview...
Overview

The Challenge

We are on a mission to find which specific open source technologies need to be developed and integrated to power voice-enabled products. We want to open up the voice product development process, to make it transparent and accessible to everyone.

The first two steps in this direction have been the Common Voice project, the compilation of a multilingual, open and publicly available dataset of labeled audio samples to be used to train voice-enabled applications, and Mozilla Speech open source projects (text-to-speech engine, and speech-to-text engine). 

Now, we’re taking the next step: we would like to identify and integrate a full plan to define the voice technology space by creating a “stack” of open source technologies to support the development of new voice-enabled products. Click on the orange ACCEPT CHALLENGE Button to apply!

 

The Objective

To create a comprehensive stack of open-source voice-enabled technologies for the development of voice-enabled products. 

We envision that some elements of this stack would be the following technologies:

  • Language parsing: stemming, entity recognition, dialog management, and summarization
  • Intent parsing
  • Keyword alignment
  • Keyword spotting
  • Voice-signal processing
  • Natural Language Processing (NLP)
  • Text-to-speech (TSS)
  • Speech-to-text (STT)

We want to improve this list by adding more relevant technologies and then place the list in a public repository for open access. To achieve this, all proposed technologies in the stack need to be open-source licensed.

 

Your Solution

When composing the stack, HeroX Solvers should answer the following questions:

  1. Which specific technologies need to be added to this list? Please, provide a detailed explanation of why each of them deserves to be added to the list. The quality of this explanation is as important as the number of proposed specific technologies.
  2. What open source options are available for each of these technologies? Please, provide an estimate of their quality.
  3. What would a full stack of open source voice technologies look like? Please, choose an open-source project for each piece and describe the “architecture” that would bring all the pieces together. Tell us why you made the choices you did.
  4. How would these open source options work with one another? Are there any technical challenges to be solved to make the stack work?
  5. Are there any “gaps” in the stack where quality open source projects are not available? If yes, please provide a description of that gap and a thorough explanation of why existing open-source projects are not useful and what functions a new project would need to perform to fill the gap.

Prizes

A total prize pool of up to $6,000 is available

  • Up to 2 grand prizes x $1,500 each
  • Up to 2 runner up prizes x $750 each
  • Up to 3 honorable mentions x $500

 

Timeline:

  • Open to submissions: June 20, 2019
  • Submission deadline: September 19, 2019
  • Judging: September 20 to October 21, 2019
  • Winners Announced: October 24, 2019

Other links:

  1. Recent Updates
  2. Forum

 

Click on the orange "Accept Challenge" Button to apply!

Guidelines
Timeline
Updates 14
Forum 2
Community 130