AIVON Protocol - What is Human Expert Network
in spite of rapid advancement in
the sector, AI (artificial intelligence) is not ideal; humans carry on to play 2
vital parts in the procedure. First, they aid label the content that'll be fed to the AI Algorithms. Then, they correct
any incorrect predictions made by the AI tools to enhance the perfection of
future results.
We'll use many disciplinary open sources network of freelancers with knowledge,
tagging, logging metadata moderating, management transcribing as well as
translation. AIVON Network empowers its
community with devices which allow individuals to make added meta data. Of crucial value is the value of content
listing by supplying perfect descriptions and assignment of appropriate tags
backed on the most largely accepted taxonomies and lexicons.
This is the place where power
tools and human expertise are needed. There're surely very diff lexicons for
medical conference recordings, sporting events news stories. Individuals will
have the option to contribute specific subject places, according to their human
expertise.
The community will also be linked
in review, moderation, and verification of desire meta data, including classifications assigned by AI, as-well-as
correcting the transcriptions as well as translations. How difficult is for
humans to right transcript generated by AI?
We've seen recent researches which
show that device translation is perfect to 95 percent for broadcast recording,
but recognizing speech from an iPhone on
a table in a cavernous conference room with 5 people discussing could be 70
percent or less. Humans will surely prefer correcting transcriptions generated by
Artificial intelligence, rather than writing from scratch.
AIVON Network gift these open
sourced human professionals with AIVON Network tokens on their perfection and productivity.
This approach not just provides a wider pool of human professionals to pick from,
the diversity of decision making procedure and the use of a consensus network can
eventually lead to super outcomes in AI perfection.
Each and every face, scene,
object identified or sentences translated or transcribed comes with confidence rate.
For confidence score below are a certain
threshold (that might vary over time and for diff attributes) or if the Artificial
Intelligence nodes are not able to detect
any meaning full context, the VALIDATOR can send these parts to the human professional
network for additional verification as well correction.
For easy verification jobs, AIVON Network will pose its
human professional questions that are
same to picture recognition re-Captcha,
meaning human professional can be offered a random figure of celebrity photos and
be asked to pick the ones that belong to
Tom Cruise for instance.
As proof of job, the human professional can be asked to
fully a same RECAPTCHA picture selection of which the outcomes have been determined
by the parent VALIDATOR node. This node could also send similar tasks to a panel of randomly picked human professionals,
the similar answers provide by the bulk
of the human professionals would be deemed the right one.
Pulling It All jointly
AIVON Networks be part of a wider
content system consisting of content advertisers, generators, and streaming
media service givers. At AIVON Network, we can see how the AIVON Network a block chain based protocol allows collaboration
and cooperation.
The AIVON Network community will give
an open, crowd-sourced and distributed community for translation and transcription.
Content holders and distributors in need of translation and transcription of
their video content will arrive into the open market to translate and transcribe
their video content through AIVON Network’s AI mechanism translation and transcription
with allocated crowd-sourced computing resources joint with a open sourced open
community of translators and transcribers, the entire incentivized using the
AVO (AIVON Network)Token.
For more details about this AIVON Network project can be found on the official site.
• Website: https://aivon.io/
• Whitepaper: https://aivon.io/download-whitepaper/
• Facebook: http://www.facebook.com/aivonio
• Telegram: http://t.me/aivonio
• Twitter: http://www.twitter.com/aivonio
• ANN Thread: https://bitcointalk.org/index.php?topic=5043237.0
#Aivon #aivonico #tokensale #AI #Blockchain #aivonio
------------------------------------------------------------------
bitcointalk profile url - https://bitcointalk.org/index.php?action=profile;u=1882134;sa=summary
bitcointalk username - araya95
ETH address - 0x2BAE99200F4b0904ed29f6DF2c5677D676614E17
Comments
Post a Comment