If someone asks you what are the most interesting technological advancements recent world is witnessing, my answer will be the following:

Advancements in Machine learning (which sprouted from AI) technology, specifically Deep learning
Blockchain related technology

What happens if you combine both? Answer: MAGIC !!!

That is what EFFECT.AI is trying to realize. Let us see how they are going to do this in the coming sections.

What is the problem EFFECT.AI trying to solve?

There are a lot of advantages which Artifical intelligence brings to the humankind. There are many problems which traditional computing is unable to solve. Say for example classifying images, face recognition or language translation whatever be it. But there are still some challenges which AI community needs to work on. Effect.AI identifies these problems as pointed out below (from Effect.AI Whitepaper):

Data processing challenge: For typical Machine learning(ML) algorithms specifically Deep learning(DL) techniques to work, you need a huge amount of annotated data. This human feedback is a tedious and costly process!
Diverging task challenge: The complexity of the tasks in applications like self-driving car ones diverges and explodes since the primary focus is to deal with steering and driving car, but there are other tasks like identifying road signs all around the world!
Computational Cost challenge: This is the most well-known issue. For the DL algorithms to work the computing power needed is enormous. Like a lot of expensive GPUs(Graphics processing units) are needed to run these algorithms. Not just cost of GPUs, immense amount of electricity is also needed to execute these algorithms.

The solutions from EFFECT.AI

Quoting from Effect.AI Whitepaper:

A decentralized ledger like the blockchain provides a direct link between supply and demand which can greatly mitigate these problems. The transparency that blockchain offers will boost discoverability on the network, resulting in a high degree of collaboration and data sharing between agents. It also increases knowledge diversity and makes AI more affordable by sharing costs. Around this vision the Effect Network is designed. Like other decentralized applications, Effect.AI directly connects supply and demand without the need for an intermediary party

Effect.AI will use the below mentioned 4 techniques to solve the problem:

By directly linking supply and demand through their micro-tasking platform Effect.AI Mechanical Turk (EMT) will make training AI algorithms easier, faster and cheaper. This will improve the accessibility of the large-scale data to the users for doing AI tasks.
Effect.AI Smart Market (ESM) is an exchange with a rich ontology of specialist AI applications where users can buy and sell information. This will enable users to choose a wide variety of applications to solve complex problems with higher accuracy.
Using the above-mentioned facilities, users will be able to solve the problems with higher performance output. Building huge dataset and training algorithms will become very easy.
By putting multiple AI algorithms on decentralized blockchain network, Effect.AI is making it possible for different AI algorithms to make talk each other. In technical language, you can call this interoperability.

The EFFECT Network

The Effect.AI project introduces an open, decentralized network that provides services in the Artificial Intelligence market. This project is called The Effect Network. The Effect Network will be deployed on NEO blockchain and used as a decentralized application as-is. But, in order for the network to grow and become sustainable, Effect.AI will be implementing a governance and the participants will have the incentive to use the EFX token for the purpose of Artificial intelligence tasks.

Governance Model: (from Whitepaper)

The three phases of THE EFFECT NETWORK

Phase 1: EFFECT Mechanical Turk
This first phase will be an arena where humans do lot of tasks which will help AI algorithms in getting trained. The incentives will be given to the humans who do these tasks in the form of NEP-5 token. This will be known as Decentralized Mechanical Turk (M-Turk).

Phase 2: EFFECT Smart Market
This is a smart marketplace where customers can buy AI services and suppliers can supply their AI products. And the payment will be in EFFECT token. Even each algorithm will have a corresponding wallet! And also think about the scenario where algorithms communicate each other.

Phase 3: EFFECT Power
This phase will run popular decentralized Deep Learning frameworks with the aid of popular libraries tensorflow, MXNet, Caffe etc. Effect.AI will distribute the computational power of all AI algorithms in the network. This will be implemented with help of projects like GOLEM which has their own decentralized supercomputing network available for access.

A typical use case

Think that you are a research student who wants to focus on your bioinformatics problem rather than spending a lot of time annotating data and waiting years to complete your task to just start asking your question. What you can do is that with the aid of EFFECT.AI you can use your research fund to put your data labeling task on the blockchain. You need not even worry about the security of your data since it is on a blockchain! And later you can even train a deep learning application for your problem to classify/predict your outcomes! Isn't that great?

A EFFECT.AI video

So why waiting? Go get the ICO tokens before March 24th

See here for the details of token sale: https://www.effect.ai/tokensale

And below is a video about ICO review:

You can find more information here:

Contest Post: https://steemit.com/cryptocurrency/@originalworks/4m2myp-300-steem-40-bonuses-sponsored-writing-contest-effect-ai

effect.ai2018

EFFECT.AI : Let decentralized network of AI help the humans invent the future!