There are two lines of development witnessed by us in the 21st century. One is the advent of Artificial Intelligence while the second is the advent of blockchain technology. Both have their own potential of disrupting the industry and current norms which we take for granted. Let’s look at both individually.
With the advent of the computer age in the past century lot of data was computerized and therefore, the analysis of such data was easier and faster. However, in nearly half a century since computers entered our lives, there has been tremendous change in the way the device itself operates and processes information. Contrasting this, the data that is captured has not changed at all – it still remains manual. This also means that the computer can only be as effective as the data captured (garbage in, garbage out); additionally, only as fast as the data captured. Work from large corporations in Americas and Europe for capturing data were outsourced to Asian countries where millions of data entry operators read through either pdf or excel or pictorial forms of data and entered them manually into the computer. This, no doubt, saved a lot of cost (cost of labor being cheap in Asia) but it still required the same amount of time or may be fractionally lesser.
Artificial Intelligence tried to solve this problem. It tried to make data capture as fast as data processing itself; in short removing manual intervention. It need not have a human operator to read, comprehend and enter data. The machine can itself sense, rationalize and capture data. This is what has spurred the aspects of Machine Learning (ML), Natural Language Processing and the likes. We now see proof of such intelligence in the form of Google Assistant, Alexa, Robotic Process Automation (RPA) to capture structured and unstructured data from pdf, excel or any other types of files. And the news is that it is working! Take a look at Sundar Pichai, CEO of Google, talking about tasks Google Assistant can do.
Impressive, isn’t it? The problem with these kinds of AI work is that it is usually centralized around the companies developing it and requires tremendous amount of money – usually in billions of dollars.
Or it can be centralized around the needs of some Governments. Still, the work of automation will be largely centralized. Take a look at this video of AI being used for individual recognition in China.
So, that’s the power of Artificial Intelligence. At the moment centralized and requires billions of dollars of investment. Now, let’s talk about the second revolution – the Blockchain.
Since the advent of Bitcoin cryptocurrency, the world has noticed few of the advantages of the crypto world. These can largely be stated in five points:
This has driven many organizations, central banking institutions and even governments to look at solutions deriving the benefits of the blockchain. Can you imagine public distribution system (PDS) of food in poor countries through a blockchain enabled network which will allow the government to have visibility right till the last individual getting the food? No fudging of information is possible (the chain being immutable) and therefore, no possibilities of corruption. How’s that for a solution?
This is what is driving blockchain solutions in many areas – financial, pharmaceutical, and government distribution to name a few.
Incidentally, when we look at the two revolutions of the 21st century – AI and Blockchain – they seem to have some divergent tenets. AI is more of a centralized approach while Blockchain is completely decentralized. AI’s scalability is driven by the company’s financial strength to invest, while blockchain’s scalability is driven by its speed of transactions and individuals planning to use it. Development of AI itself is a secret affair and only the end product is witnessed by the users whereas blockchain is transparent and visible to all touch points at all points of creation.
Completely contrasting solutions yet, one can’t stop thinking what if these two combined and brought in the advantages of each other to an end user? If that happens then are we talking about twice the revolution? Is it even possible? Has anyone given such a solution a thought?
It appears someone has.
Effect Network, in my opinion, is the one which bridges the gap between AI and Blockchain and gives the concept of development and testing a whole new meaning in a decentralized environment. Let’s get to know them in their own words.
“Effect Network introduces an open, decentralized network that provides services in the Artificial Intelligence market.”
“Effect Network runs seamlessly on NEO blockchain and is fuelled by a network NEP-5 token.”
Think about it this way. Here is an AI developer for a particular use case but does not have enough test cases for the AI bot to learn – what will he/she do? Two things – hire people to create test cases for the bots to learn OR seek the help of users on the net to interact with the AI bot so that it can learn. If you are a Google then the first plan works for you, however for most, the second would be the plausible workable solution, albeit a small fee. That is what Effect Network achieves. How’s that for decentralizing AI learning?
I am sure most readers would concur that the concept is quite strong but how does it work? Well, Effect Network is planning the launch of the whole artificial intelligence concept in three phases. Here’s how.
In this phase Effect Network brings together the creators of Artificial Intelligence on one end and the testers and users on the other end. There are two distinct sets of work that can happen. On the base AI coding that the creators have done, they may ask developers for some coding help, as much as ask the end users to test or help the AI bot to learn. So, work that the creator requires can be of two distinct nature – Coding as much as Assisted Teaching. This of course, in return for a fee that is paid to the individual performing the task with the network NEP – 5 tokens.
In short, in this phase the creators, users and coders are brought together. The coolest part is that this phase is going to be launched soon, very soon! Hear the team speak over here.
This is the second phase where buyers wanting to buy the AI solution or Algorithm can do so from the seller or creator of the AI solution. For example, if a creator has made an AI solution for speech recognition to identify the language of a person speaking, it could be a solution for a telecom company which may get callers speaking in different languages. So, you don’t have to spend endless amount of time on the IVR (which by the way plays in English and 2 more languages at the most) punching numbers indicating which language you want to speak. The AI solution will hear the caller as soon as they connect, identify the language, and start responding in that language or route the call to the person who knows the language. Isn’t that cool?
So, the telecom company in our example can buy from the creator and deploy it on their IVR – An automated IVR. And if there is no exclusive contract then the seller can sell to many telecom companies and make some cool money. Also, I bet the solution would be cheaper for the telecom company to buy from such a market than from a Google. Wouldn’t you agree? So, this platform inherently makes the end product cheaper for the buyer.
I guess this phase is quite self-explanatory. When we speak about AI, the computational power required is phenomenal. The more the learning the more the power required. Now, coming from one computer or a data-center in one location, the power requirement will have a cascading effect on the cost. Depending on the cost per unit of power in the country where the work is being done, the overall cost can range from absurd to prohibitive. So, what is the solution?
Now, isn’t the blockchain decentralized? Yes. Then can the power generation be also shared from different ecosystem of computers across the globe? Yes, again. That’s what Effect Network intends to do in this phase. In fact, Effect Network plans to partner with Golem for cost-effective computing power or create the framework of smart and cheap computing power usage themselves.
Phase 2 and Phase 3 will be launched soon after Phase 1 launch.
Well, the answer to this lies in analyzing the solution itself, the worker community and the cost of solution to the end users and the world. Let’s chart out some of them.
1. There are only few companies like Google, Amazon, Volvo (driverless driving) and others working excessively on AI at the moment. How many of us work with them? Not many. How many of us understand AI? Probably a lot more than those employed by these companies. Effect Network tries to reach out to such workers who understand AI and enable them to contribute while making money. This is what I call Decentralized Job Creation!
2. All the companies that we just spoke about will look at automation in their sphere of work. So, mobile automation, car automation, etc. may follow but what about other areas? Do we wait till some big company takes it up? Can the serving of food in a restaurant be automated thereby reducing wait times for customers? The answer is yes, a restaurant in Japan does it but may not interest a Google or a Volvo. So, automation ideas beyond the big companies need a conduit to express themselves and find a market. This is where Effect Network allows small time AI developers to present world changing solution to the market. Our earlier discussion of voice recognition IVR being sold to a telecom company is a good example. This is what I call Equal Opportunity Creation.
3. The AI learning cycle itself would be extremely divergent and faster. Isn’t it obvious when a bot which needs assisted learning is being taught by users across 200 odd countries of the world, won’t the bot learn more and learn fast? This is a single most important factor for which, I suspect soon, Google, Facebook, Amazon and the likes may turn to Effect Network for help. This is something not even these companies can achieve. We can call this Decentralized and Faster Assisted Learning.
4. Idea curation and Idea streamlining would be faster over here. After all, the phase 1 is all about bringing creators in touch with other developers, wouldn’t there be channelization and discussion over ideas? Wouldn’t that bring about improvements at a much faster rate? So, Effect Network will help crunching "Ideation to Completion" phase of AI development; not to mention ironing out issues and bugs at much earlier stage and faster rate. This is nothing but speeding up AI Development.
5. We touched upon the point that the AI solutions built and sold in the Effect Network ecosystem will be much cheaper than those coming from multi-billion dollar organizations and hence the cost of solution becomes cheaper for the buyer and in effect for the economy and finally for the world. This is nothing but cost-effective AI.
So, these were some points which without much effort comes forth as advantages of the Effect Network world. There are many more and will, no doubt, be unearthed by the community very soon.
It is also recommended that readers understand more about Effect Network through any of the following informative resources.
This article is written in response to the originalworks’ call on authors’ thoughts on Effect.AI as a concept. It can be read here.
Image Courtesy: Geralt at Pixabay
Disclaimer: The videos added to the article belong to the users who have posted them on youtube. The Author does not own them nor validates that they belong to the ones who posted them on youtube. The videos are included to give some additional information about the subject being discussed.
effect2018
