In the past five years, we have witnessed a rapid growth in the number of Artificial Intelligence (AI) algorithms with practical applications. Smart services, like self-driving cars, face and voice recognition on mobile phones, and image translation are getting a central place in everyday life.
Even though the AI market is growing at a remarkable rate, it’s becoming less accessible with each passing moment. The main reason for this barrier is the control of large corporations on the majority of AI developments.
Effect.AI proposes a solution for this barrier by creating an open, decentralized network that provides most, if not all of the services required for an accessible Artificial Intelligence market. This project is called The Effect Network. The Effect Network ensures that the platform will be open, easy to use, and scalable. It requires no commissions, has a low barrier to entry, and provides fast growth for its users.
Effect.AI aims to accomplish this mission by deploying the project on three platforms that run on the NEO blockchain and are fueled by a network token called EFX.
With a projected size of 15.7 trillion dollars as early as 2030, the Artificial Intelligence market is on track to become one of the largest and most important markets in the world. Ubiquitous mobile supercomputing, intelligent robots, self-driving cars, neuro-technological brain enhancements, and genetic editing are just examples of major developments in this revolution.
Artificial intelligence is intelligence presented by machines, i.e., a computer program that is programming itself to do its task based on the inputs it receives, just like a child learns how to name objects by his mother. Unlike regular computer programs, which humans need to write code, so they can perform their tasks.
An example of AI learning used by Google would be, google photos search engine.
With Google Photos, you can search for photos containing objects, animals, certain people, and more. Google photos is one of many examples of an AI algorithm trained to do a specific task. In general, if you show the AI algorithm a million photos of cats, it will eventually learn how to recognize a cat by itself. The same way of training is done to any specific object in the picture.
Some people consider that AI technology is an industrial revolution. For the revolution to happen, man and machine cooperation are sometimes necessary.
The most efficient way to train an AI algorithm is by letting the AI program interact with another program- a self-taught algorithm. We recently heard of the AlphaGo Zero, a self-taught bot, created by DeepMind (a company owned by Google), that recently became the best Go game player in the world in just 40 days. Go game is known as a very complex ancient Chinese board game.
AlphaGo Zero was only taught the rules of the game, it learned how to play by playing against itself, using random moves. After 40 days of training and 30 million games, the AI was able to beat the world’s previous best ‘player’ — another DeepMind AI known as AlphaGo Master. The results were published in Nature, with an accompanying commentary.
Though the self-taught AI algorithm is the most efficient way of training, sometimes human input is also needed. For example, I wish to create an algorithm to recognize the name of any flower in a picture. I could use millions of pictures of various flowers and ask people to name the flower in each picture. The AI algorithm will eventually be trained to successfully name a flower in a picture.
Currently, if a developer wants to train their algorithm using humans for a specific task, one of his options is to log in to the Amazon Mechanical Turk. There, he puts in a request for specific work to be done for his algorithm. His request is added to the request list. Workers, on the other side, can apply to complete the requester’s work from that list.
On its platform, Amazon is currently transacting around 500k tasks each day. Amazon is charging around 20% on transaction volume.
Effect.AI is aiming to disrupt the Amazon Mechanical Turk market by utilizing smart contracts deployed on Neo blockchain technology to cut out the middleman and by doing so lower the fees substantially. Moreover, blockchain transfer of value is not subject to the fees involved in the current banking system.
Effect.AI proposes a private and decentralized ecosystem for AI development and AI-related services. This network will be called The Effect Network.
Because of the size of this undertaking, The Effect Network will be deployed in three consecutive phases (see below), allowing the healthy progression of development and adoption. All of the phases are fueled by the EFX utility token.
A private, decentralized, marketplace for tasks that requires human intelligence. This phase gives businesses, and aspiring AI developers much needed access to a large workforce of human intelligence, and it’s based on centralized business models like the Amazon Mechanical Turk. A key difference is that the Effect Mechanical Turk is a peer to peer, meaning supply and demand are connected directly and more efficiently. The crowdsourcing technology enables requesters to create tasks to be completed by workers in exchange for compensation. The job offers are called Human Intelligence Tasks or HIT’s for short.
The provider of the HIT’s are called Requesters. These HIT’s are created by anyone who needs the power of human intelligence to perform tasks that computers are currently unable to do. These tasks are described and compiled through smart contracts on the blockchain.
The Requesters can register their tasks on Mechanical Turk for completion by Workers. The Requesters decide how much EFX the Workers get for each completed task. In this way, the Effect Mechanical Turk gives Requesters access to an on-demand, scalable, and distributed workforce.
It’s important to note that the platform will exclude or include workers from tasks based on certain criteria that were chosen by the requesters such as location, age, gender, etc.
Workers can accept tasks from Requesters at any time, anywhere, and from any device. When a Worker completes a HIT, they are paid with cryptographic EFX tokens.
The Process of submitting tasks to the Effect Mechanical Turk
The second phase is a marketplace connecting app owners seeking an algorithm to teach their program a specific task and AI algorithms proven and built by developers.
For example say I’m an online store owner, and I would like to develop a computer program to track my customer’s behavior in my store. Then based on that I want to offer them the right products with the highest yield, I could use the future Effect Smart Market and purchase the right AI algorithm for my needs.
Following their data interchange format (and by specifying a usage fee for consumers) an application owner can register on the exchange by specifying a public endpoint for his or her application. This application can now be invoked through smart contracts on the blockchain. The caller of the contract will have to transfer the required funds to the owner of the contract to get an authorization token that allows him or her to interact with the application.
The exchange protocol can be built directly onto the Effect.AI interface where the agents receiving EFX tokens are the ones supplying AI algorithms, and the agents providing EFX tokens for these services are the Requesters. The Effect.AI Galaxy Pool performs its role in the background to assure liquidity (for more details about the Effect Galaxy Pool see below). For example, we can take an agent that has a combined system of 4 AI parts and has one AI part missing; the existing AI network can improve the availability of the missing AI part.
Phase 3- Effect Power
The first two phases of the Network include decentralized data gathering and the use of AI algorithms. Up to this point, the algorithms themselves still run on centralized servers. In the final phase of the network, the actual computation will be distributed so that the algorithms run globally without a single point of failure. To achieve this, The Effect Network uses the fact that most machine learning algorithms have a rigid structure and operate on sets of weights. These types of algorithms are relatively easy to distribute.
The Effect Network decentralized compute engine is based on popular Deep Learning (DL) networks like Caffe15, MXNet16, and TensorFlow17 where the network structure can be defined as a declarative graph and weights are stored as matrices of real numbers. These matrices can be distributed over a decentralized file system and processed at different compute nodes on the network. More detail on this phase will be provided as the project progresses.
The Galaxy Pool is a central pool of EFX, and native tokens built to ensure the necessary level of liquidity is always maintained under several rules. This allows:
This kind of liquidity can be hard to achieve for a new token on the market and can be hurt by speculative trading. Therefore, the Effect Network will maintain this pool of tokens to provide liquidity, encourage adoption and stabilize network fees. The Pool will consist of a mix of EFX and native tokens to function correctly, and as NEO is indivisible, the rate should be defined in EFX/NEO.
One of the important basics of the system, which is also connected to the Galaxy Pool, is the invention of the G-EFX Token along with the EFX token.
The key to achieving stable exchange rates for users of the network, at all times, is making a distinction between G-EFX tokens and EFX tokens. G-EFX tokens can be bought, but any G-EFX bought from the Pool cannot be sold back to the Pool. A G-EFX token is cleared (converted to a regular EFX token) by spending it through an Effect.AI Service Contract. These are the service contracts from the tasks and service registry.
This protects the Galaxy Pool from external manipulation and keeps exchange rates stable for all agents in the network. Furthermore, to assure Requesters do not have an overflow, G-EFX will gradually expire and return to the Galaxy Pool over time. The Pool compensates the Requester for this by offering a favorable exchange rate for the tokens used to purchase G-EFX.
The team has been working hard on the project for the last 8 months. The team currently consists of 15 full-time developers.
Chris Dawe – Co-founder and CEO, with a diploma in Marketing and International Business from Kingstone international college (Australia) and an advanced diploma in leadership and management from Danford College (Australia), Chris is an entrepreneur with experience in Business around the world. Former CEO of Choice Energy in Canada for five years.
Jesse Eisses – Co-Founder and CTO, AI project lead. Jesse owns a Bachelor’s and Master’s degree in Artificial intelligence from the University of Amsterdam.
Nick Vogel – Co-Founder and CCO, UX project lead. Nick has more than ten years experience of in media, web, and marketing design.
Charlie Shrem – Charlie is the founder of the Bitcoin Foundation, which has allowed partnerships and mergers of some of the local Bitcoin communities around the world to have global resources and mentoring.
Tony Tran – Tony Yang received his B.S and M.S in Computer Science in 2007 and 2009 respectively. He has experience working with big corporations like Uber and Cisco as a software and data engineer and recently founded a blockchain platform called the Bee token.
All team members and advisors are listed on the Effect.AI website
The EFX token is the only mean of exchange between parties in all phases of the Effect network.
Done: MVP of the Mechanical Turk marketplace.
Q2 2018: Effect Mechanical Turk (Beta on Neo) – phase 1.
Q1 2019: Effect Smart Market (Platform release) – phase 2.
Q4 2019: Effect Power (Framework release) – phase 3.
Token type: Neo Nep5 token
Hard cap: 14.8 Million Euro (~18 Million USD).
Token allocation: Finite number of tokens – 650 Million EFX. 40% distributed to public, 20% secured for potential future development, 16% smart galaxy pool, 15% team (locked for 2 years), 7% partnerships/advisory, 2% bounty program.
Private sale: Small seed funding was raised for the launch of the ICO.
Crowd-sale (14.8 Million Euro): Fair share distribution, a potential max individual cap of 25,000 EURO, No lock up. 10% bonus for the first 5.2 Million EFX.
*The 20% of tokens for future funding will be locked by a smart contract for 18 months. If predetermined milestones are met, these tokens will be released for the second round of funding for the later phases of the project. If predetermined milestones are not met the tokens will be burned.
Almost all of the funding of ~18 Million USD is to be raised via public sale. Considering the over-whelming community interest, 25,000+ Telegram members to this project, the individual cap is expected to be small (~18 M USD/Number of approved whitelist subscribers).
The Effect.AI token sale model (Similar to 0xProtocol token-sale model) will create a situation where there will be almost no token holders with a large proportion of the tokens. All the above should help the token price to remain stable or rise once it reaches exchanges because of the following: 1. Members that wish to buy a larger amount of EFX will only be able to do so by buying EFX once the token reaches the market. 2. In the beginning, no whale should be able to manipulate the price on the selling side by putting a large amount of EFX up for sale.
Although the Effect Smart Market (phase 2) is to be launched in Q1 2019, there currently aren’t any major platforms providing this service. Considering the ongoing penetration of AI technology into our daily life, there should be a need for a marketplace to connect AI algorithms with any entity that is seeking to utilize algorithms with learning abilities in their company. The relative first-mover advantage of the Effect Smart Market should help the project to penetrate the market and to attract a user base.
The Effect.AI first phase is the Effect Mechanical Turk, the task of attracting users to a new platform is a great challenge. Especially when you consider that the big companies like Amazon, Fiverr, and others dominate those marketplaces.
We should expect announcements of partnerships in the near future as funding is being reserved for that purpose. On the supplier’s side (Workers) – mainly third world entities like banks and governments and on the demand side (Requesters) with AI program developing companies.
Effect.AI website
Effect.AI Telegram Group
Disclaimer: All ratings published on CryptoPotato are for informational purposes only. CryptoPotato does not provide investment forecasts, recommendations or consulting. Thus, neither CryptoPotato nor its writers can be held responsible for its visitors’ investment decisions. CryptoPotato is not affiliated with and does not endorse or sponsor any listed token sales. CryptoPotato may receive a fee for advertising certain token sales, in which case it will be designated and marked accordingly. CryptoPotato may receive a fee for conducting due diligence before listing, rating and/or evaluating a token sale. CryptoPotato will never accept a fee in exchange for a certain rating score and/or evaluation for token sales. Full disclaimer
