Mycroft is named after a fictional computer in the science fiction novel "The Moon Is a Harsh Mistress", published in 1966.

In Kansas City, Missouri, at a makerspace, Ryan Sipes and Joshua Montgomery discovered an smart virtual assistant project that gave them the idea for Mycroft. They were intrigued by the technology but disliked how rigid it was. Montgomery has pledged that Mycroft will uphold users' privacy rights through its open-source machine learning platform because she thinks the burgeoning market for intelligent personal assistance raises privacy concerns for users.

In 2016, Mycroft AI, Inc. won the prestigious Techweek's KC Launch competition, among other honours. During this time, the business accepted a strategic investment from Jaguar Land Rover. Before deciding to use StartEngine, an equity crowdfunding platform, to sell shares of the company to the public, the company had already raised more than $2.5 million from institutional investors.

Mycroft AI is a voice-activated virtual assistant that can search the web for specific information, download a podcast, etc., in response to verbal requests. It's a great open-source software that offers privacy and platform flexibility, unlike similar software from companies that profit from personal harvesting data.

Mycroft AI is a Python program that is portable across numerous computer systems. The well-known Raspberry Pi board is often used (though not exclusively) to power the virtual assistant's voice interface. Mycroft helpfully offers a Picroft image that we can load onto a Raspberry Pi. While Picroft is an excellent option, it has some restrictions, such as not being 64-bit compatible.

Applications

Mycroft offers free software for the majority of the voice stack.

  • Mycroft's Precise Wake Word engine performs Wake Word spotting, also known as keyword spotting. Mycroft used PocketSphinx before Precise became the default Wake Word engine. Instead of relying on phoneme recognition, Precise employs a trained recurrent neural network to differentiate between sounds that are and aren't Wake Words.
  • Mycroft collaborates with Mozilla's Common Voice Project to use DeepSpeech speech-to-text software.
  • Mycroft employs an intent parser called Adapt. Adapt parses intent by matching specific keywords in a particular order within an utterance. Padatious, their parser, is also available. Padatious, on the other hand, determines intent through example-based inference.
  • Mycroft employs Mimic, based on the Festival Lite speech synthesis system, for speech synthesis.
  • Mycroft is designed to be modular, allowing users to swap out components. For example, instead of Mimic, espeak can be used.

Hardware

Mycroft is also developing and selling smart speakers that run its software. Its entire hardware is open-source and distributed under the CERN Open Hardware Licence. Its first hardware project, the Mark I, was aimed primarily at developers. The developers partially funded its production through a successful Kickstarter campaign. Units began shipping in April 2016.

The Mark II, a piece of hardware it recently finished, is aimed at users other than developers. The Mark II, in contrast to the Mark I, has a screen and can communicate information visually and audibly. Similar to the Mark I, the Mark II's production was partially aided by a Kickstarter campaign that ended in February 2018 and raised nearly eight times its initial funding target. 

Conclusion

Mycroft has engaged in several business partnerships. The company carried out bulk machine learning training in May 2018 in collaboration with WorkAround, an impact sourcing provider facilitating employment opportunities for refugees. To determine the prevalence of coughing on public transportation, Mycroft worked with the disease surveillance and forecasting firm SickWeather in October 2018. This project was funded by the City of Kansas City, Missouri.

Want to publish your content?

Publish an article and share your insights to the world.

ALSO EXPLORE

DISCLAIMER

The information provided on this page has been procured through secondary sources. In case you would like to suggest any update, please write to us at support.ai@mail.nasscom.in