laurayco / Jarvis

Exactly what you want it to be.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Jarvis

What is Jarvis?

Jarvis is an operating system project aimed to be an AI based personal assistant capable of dynamic learning and knowledge growth backed by Wikipedia and various repositories of web data.

The issue with modern / current AI based personal assistants (ie: Google Now / Siri / etc. ) is that their functionality is limited to various pre-defined keywords. The Jarvis project aims to enable dynamic / fuzzy natural language processing to disambiguate various commands, queries, and declarations (ie: "Play 'Back in Black'" vs "Turn on 'Back in Black' or even, "Replay that last album on high volume). It is unlikely that a software developer can anticipate every possible configuration of conditions an end user may specify, so the solution is to create a software capable of identifying different, sometimes ambiguous, conditions / commands / etc.

Some immediate planned features for Jarvis (core functionality):

  • Specific tasks / jobs at scheduled interviews ( think chron jobs )
  • Maintain a calendar, with intelligent reminders ( travel plans based on historic traffic data, for example )
  • Communication features
  • sending emails, text messages, initiating phone / skype calls.
  • Query data (think Google "card results")
    • "How old is Barack Obama?"
    • Who is currently the prime minister of Australia?
  • News synthesis

It also is desirable that this OS be portable and have an installation fit on a modest storage space (think 16GB minimum, 32GB average, 64GB ideal). Anything larger and a cloud based service becomes more feasible for the sake of portability.

In-home integration is also desired. Potential integrations include Nest systems, for example. The system should be able to control these systems with a "zero configuration" method, where the AI intelligently predicts / detects home presence, identities of persons, etc.

About

Exactly what you want it to be.


Languages

Language:JavaScript 44.6%Language:Python 30.1%Language:C++ 18.2%Language:Makefile 4.1%Language:HTML 3.0%