AIVON - Artifical Intelligence Video Open Network
AIVON Network is a new decentralized protocol that's created and made via the power of an AI system. Owners and innovators of the AIVON Network program put jointly research and hard work to solves many problems in video files meta data, the safety of content, and internet video file localization. The AIVON Network will apply AI to make sure the safety of data from violence, nudity, alcohol, adult, guns and offensive language to name a few. The AIVON Network will generate a confidence rate and turn it into a content graph for accurate demonstration and viewing by consumers.
AIVON is a decentralized, open-source blockchain protocol and ecosystem being built on a consensus network of Artificial Intelligence (AI) computing 10 resources and a community of human experts, working together to generate normalized and enhanced metadata for video content. This enhanced metadata will be the basis for the AIVON protocol’s first decentralized application – an open search engine for video, one that is decentralized, indexed and maintained by our community. AIVON focuses on decentralizing and democratizing online video, giving power back to Internet users.
The algorithms are deployed via mining nodes focus with filtering miners video files, CPU and GPU resources. It'll also add video file transcripts, time-coded tags, and translations to index wanted and special media particularly content snippets.
In addition, this will help AIVON achieve broad adoption and network effects benefiting all participants. The blockchain implementation will achieve this with smart contracts between the content owners, advertisers, distributors and service providers. The smart contracts will maximize utilization by enabling dispatch of jobs to the most productive service providers and grant incentives to promote better quality.
AI Computer Vision (CV) algorithms running on nodes using CPU/GPU resources will be used to scan media files, generate enhanced metadata including time-coded tags, classification, categories, transcripts and translations, and an index of the video objects. Humans with expertise in tagging, editing, moderation, transcription and/or translation can participate in the AIVON shared economy to help with the verification, validation and/or creation of video metadata. AI Machine Learning (ML) algorithms will continually learn from the actions of the AIVON community to become better and smarter. AIVON will empower the community with tools to enable the community to moderate, review, verify the meta-tags, categorize, transcribe and/or translate content, as well as provide economic incentives to encourage this activity through AVO Tokens.
- Artificial Intelligence
Specialized AI algorithms will be deployed on mining nodes, so that miners’ CPU and GPU resources can be used to scan media files, generate the enhanced metadata including time-coded tags, classification, categories, transcripts and translations, and an index of the video objects.
- Content Graph
AIVON platform will use AI to define a confidence score for each of several content safety attributes, such as: nudity, adult, offensive language, hate speech, violence, guns, alcohol etc. These confidence scores will be combined into a vector called a Content Graph which can be visualized as a bar graph.
- Human Expert
AIVON uses a network of freelancers with skills in tagging, metadata management, transcribing and translation. AIVON will empower its community with tools which enable individuals to review, verify and correct the AI-derived metadata, including the categorizations, transcriptions and translations.
They are five characters in all;
- Human Experts (Same to motorized Turks)
- Computing Resource Providers or CRP (nodes running the AI)
- Software developers
- Requesters submit videos files to AIVON Network to obtain a normalized meta data of videos file.
- Requesters also pay for the job in the resident token of AIVON Network (AVN).
CRP or Computing Resource Providers
- Providers run AIVON Network decentralized public source AI tool that runs inside isolated DOCKER container and is backed on well-liked Deep Learning frame works like TensorFlow and Caffe; Convolutional Neural Network will be applied as it's the present up to date model architecture for picture classification tasks.
- Providers share its PC's computing resource (CPU and GPU) to run AIVON Network AI tool to generate the improved metadata including Content Graph.
- Providers earn AVN token in exchange for the given computing resource and jobs performed.
HE or Human Experts
- Human Experts also give further meta data that can't be simply developed by AI
- Human Experts perform easy micro tasks to authenticate that the productivity from AI.
- Relevant outcomes can be fed into the Artificial Intelligence network to teach and enhance it.
- Human specialists get token in exchange for finely performed tasks.
- Validators stake a large amount of AIVON Network Token and earn the rights to perform the following jobs in accordance to an agreement: (i) Picking-up the chunks of video content jobs and distributing them to workers and providers. (ii) Validation of the outcomes submitted from workers and providers (iii) Analyzing and combining the outcomes to produce a solo coherent Content Graph for every video.
- Validators mint fresh block-chain on AIVON Network plasma, earn a token for the task and assign rewards to contribute CSPs and is.
- Validator might lose a part of the staked tokens if it's not performing the jobs in accordance to an agreement.
- Validator decides the upcoming validator through pseudo random algorithm that's separately provable and bias-free.
- IVS will be developing at least the following two apps: a. Open Source SE and related API b. Ad matching service and related API.
- Developers play a main part in the AIVON Network ecosystem by developing apps that interact with AIVON Network or utilize Content Graph.
- Developers earn access to a list of non-confidential videos files processed by AIVON Network, with their corresponding Content Graph.
- Developers would be capable to utilize AIVON Network to make new apps which they'd charge using their tokens.
- Requester passes a video to be processed through a client software.
- The client software cuts the video up into individual chunks of pre-defined duration and submit them into AIVON
- Using a Proof-of-Stake algorithm, validators are randomly selected with probability proportional to their stake. Selected validators then pick up a set of chunks from the job pool, duplicate them multiple times, and add in random challenger “markers” unique to each duplicate copy at random sequence to create temp clip.
- Validator assign random AI nodes to process each clip. Upon result submission by AI nodes, Validator will use the random challenger marker to verify the result as a Proof-of-Work.
- Results are submitted to human experts to perform a binary verification.
- Results from human experts are gathered and processed by Validator.
- Requester monitors the network, once all tasks for chunks that make up a complete video are completed, Requester would be able to make a claim by providing the order of the chunks and publish an aggregated score of the original video, also known as Content Graph.
- Claim by Requester will be evaluated based on consensus. If it passes, Validator will publish the result as Content Graph onto a new block on AIVON Plasma Chain.
Aivon is being developed on a hybrid blockchain foundation consisting of the Ethereum Mainnet used to support its native AVO token and the Plasma Network that will underpin the core platform. To classify their videos and receive metadata, content providers would need to upload them on the platform and offer an AVO reward pool, optionally making use of Aivon's token credit program. Each video will be divided into clips that will be assigned unique identities and analysed by a convolutional neural network consisting of computer vision and machine learning algorithms.
This network will be distributed between AI Nodes being ran by ordinary users in exchange for a share of the AVO rewards. When necessary, the algorithms will be assisted by a network of freelance human experts, who will help with such tasks as transcriptions, translations, preliminary verifications and assigning more complex types of metadata and will receive AVO based on the quantity and quality of their contributions. Users could also stake AVO to run Validator Nodes, which will coordinate the network, assign tasks to human and AI participants, verify the end results and ensure the distribution of rewards through a proof-of-stake algorithm. They will also receive rewards but risk losing some of their staked tokens if they end up going against the final consensus.
The metadata gathered by Aivon will include video file properties, text, speech and face recognition and tags related to the video as a whole or some specific part of it. Once the analysis is complete, the original uploader could claim the reassembled video by providing the clip sequence and, if the claim accepted by the validators, receive its metadata. Aivon will publish a ContentGraph with each uploaded video, rating how appropriate it may be on the basis of such factors as adult themes, nudity, violence or offensive language. Software developers could then use Aivon and its Content-Graphs for their own video-related dApps. iVideoSmart intends to lead the way with an Open Video Search Engine (a metadata crawler governed by community consensus) and an ad matching dApp that may enable such innovative in-video advertising methods as digital product placement, ad overlays and interactive touch hotspots. AVO would, of course, be usable in those dApps as well.
AIVON Block-chain protocol releases ERC 20 tokens or also known as AVO token to incentivize and reward their members or consumers. The said token will also be applied as a type of payments in the entire AIVON system. The said podium has an Initial Coin Offering token supply of 350,000,000 and a hard cap mass of $18,000,000 USD. AIVON accepts currencies in either ETH or USD. The cost of one AVO is equivalent to = $0.16 USD.
- Token Symbol : AVO
- Private Sales Start Date : 1 August 2018
- Whitelisting Date : 1 October 2018
- Public Sales Date : November 2018 (TBC)
- Initial issue size : 400,000,000 (400 million)
- Maximum issue size : 1,000,000,000 (1 billion, after 10 years)
- Token Price : 6.25 AIVON to 1 USD (Token price in Ether will be fixed 48 hours before start of the sale)
- Hardcap Target : USD18,000,000
The public will be offered 35% through the token sale, and the team (including founders and early backers of the project) will get 20% with 7.5% going to advisors. 22.5% will be set aside as for community development initiatives, while 15% will be reserved for future use.
For more information, please visit:
ANN Thread: https://bitcointalk.org/index.php?topic=5043237.0
Bitcointalk profile: https://bitcointalk.org/index.php?action=profile;u=1847143
Ethereum Address: 0x964c905b602f3dcff74c20525184fa87021c17a9