Effect.AI; The Panacea To Major Problems Revolving Round AI Development and The Opportunities It Creates.

THE GLITCH Of Artificial Intelligence.

Artificial Intelligence (AI) is the concept of computer systems that are capable of performing tasks that usually only human can do, such as decision-making, visual perception, speech recognition and so on. source

pixabay

Despite how useful these machines are or meant to be, people are being faced with myriapods of problems as regards the research and development of the Artificial Intelligence (AI) technology. These constraints could divided into three as highlighted below;

Data Processing

It is highly inevitable to have a working or functional AI without any form of data programmed in it. This means artificial intelligence needs data to know more about the outside world at large, but the amount of data needed by these machines is often overlooked.
Neil Lawrence says;

These systems don't just require more information than humans to understand concepts or recognize features, they need thousands times more.

Therefore, feeding them with the tremendous amount of data probably needed is not in any way easy for the developers. Even if the data is provided, there will still be some shortcomings attached to the functioning of the system.

I could remember vividly, three years ago, while I was using one of the infinix products; Infinix hot 1. Then, I wanted to make the phone more secured, so I switched from the pattern mode of security to the face recognition feature of the phone. This seemed to be working perfectly well at first, but it got to a point where I noticed this feature started malfunctioning and the security of my phone got breached that people easily had access to it. Even sometimes I tried unlocking the phone myself using this feature but would only get a response that; face not recognized!

This actually led me into a research so as to know where the problem lies. I later found out that the malfunctioning of this feature could be due to limited amount of data the feature is provided with.

In most cases, an inexpensive phone’s facial recognition relies on just the front-facing camera and some not-so-advanced algorithms, maybe using a flash to take better photos.

So, getting enough data to feed them with is very demanding task.

Multitasking or Diverging of tasks

This is another major problem AI are subjected to, because to be able to multitask, a lot of good work has to be done.

Let's use Google's DeepMind, which is at the leading position of researching AI as a case study here.Three years ago, Google's DeepMind announced it had built a system that could thrash 49 Atari games and this was actually a great achievement, but each time this system beats a game, it had to be retrained for the subsequent games.
This actually made Raia Hadsell to say;

"You can't try to learn all different games at once, as the rules end up interfering with one another. You can learn them at a time, but you end up forgetting whatever you knew about the previous games. To get to artificial general intelligence, we need something that can learn multitasks, but we can't even learn multiple games."

Computational costs

The computational costs of AIalgorithms is the major problems in the development of the AI technology as regards the project phases involved. The phases include;

• Discovery and analysis phase; which is the phase of feasibility study and it is where the project objectives are being set up.
• Prototype implementation and evaluation phase; is a very crucial phase which is mainly adopted for the estimation of costs of AI development. The prototype is used for testing feasibility and proof of concept.
• Minimum Viable Product (MVP); as the name implies, is a real product with a set of essential and working features which are developed based on the findings obtained from the prototype.
• Product release; After a kind of verification has been made as regards the viability of the products, then, it can be released for use by the targeted users.

In each of these phases, a huge amount of money has to be incurred for an efficient product, and this subsequently make the the computational costs relatively high.

Who are to provide ultimate solutions to the aforementioned problems?

No matter how big a problem is, surely, there will be a solution to subdue or curb it completely. However, Effect.AI and the team is here to proffer long-lasting solutions to these problems. You might begin to wonder about what this means and wanting to know the solutions it's gonna proffer, right? Surely, I'll tell you about it.

Good News about Effect.AI & the new Effect Force (Decentralized Mechanical Turk

Effect.AI just like the blockchain, it is decentralized and provide a direct interaction between two ends without the interaction of any intermediaries or middle entities. These two ends include the the supply and the demand which can curb these aforementioned problems. EFFECT. AI adopt adopts an important feature of the blockchain, transparency. This would bring about the discovery of new ideas, collaboration and data sharing on the network which subsequently leads to integration. It will also help to bring about diversity of knowledge and make Artificial intelligence affordable by distributing costs.
Also, it will be using the idea of micro-tasking to accomplish it's goals.

The emergence of Effect Network is based on this vision. However, this network will provide; Accessibility, Accuracy, Efficient Performance and Interoperbility without and restrictions.

However, for the development of Effect Network, three different phases will be put into use. These are;

• Phase I;

Effect.AI Mechanical Turk, also known as EMT; just like some centralized freelancing platforms such as upwork and fiverr, EMT provides a marketplace where requesters and workers will be able to meet. *Requesters, just like clients, make some tasks which require Human intelligence available for Workers and in turn, the workers are compensated for the work done. The tasks provided are called Human Intelligence tasks. The provider of the task(s) called a requester while the one that executes the it is called a worker and he his paid with cryptographic EFX tokens after the completion of the task(s).

Workers can also be rated using some criteria set up on the platform. This in turn earn such worker a certain degree of reputation. Workers with high degree of reputation are able to get more tasks easily and vice versa.

• Phase II

Effect. AI smart market; The smart effect's market is a decentralized marketplace which is accessible to virtually all people from different parts of the world and various devices, where they offer and buy AI services and algorithms. At this point, Artificial and Human intelligence will be commercialized.

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 (EGP) is merely used in the background to assure liquidity. The EGP enables payments for these services with native tokens, as well as EFX tokens. At this point Smart Contracts (SCs) will ensure and improve availabilty of the services that are requested and/or provided by agents in the network. Two important pillars in the ESM are the application registry and the endpoints. The following subsections will describe the application registry and enpoints.

(excerpted from effect.ai whitepaper)

• Phase III

Effect.AI power; This could be referred to as the final phase of the network where the actual computation will be distributed and the algorithms will run globally; this will be done in such a way that failures will be greatly avoided from the system. However, this will be based on the fact that most machine learning algorithms have rigid structure and operate on sets of weight as these types of algorithms are easily distributed.

The Effect Network decentralized compute engine is based on popular Deep Learning (DL) frameworks 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 be processed at different compute nodes on the network.

(excerpted from effect.ai whitepaper)

The Role Of Effect Force In The Labour Market Of Developing Countries.

Pixabay
Unemployment in developing countries increases almost everyday at an alarming rate. Despite the increament, no certain solution has been preferred. Nigeria as a case study; In the last quarter of 2016, the unemployment rate was pegged at 14.2% by the National Bureau of Statistics (NBS) up from 13.9% in the preceding quarter. This shows a large increment in the level of unemployment.

However, the project Effect.AI is launching very soon could serve as a saving grace to the set of people in this category. The accessibility of the platform from anywhere in the world will actually go a long way in helping people secure a place on the platform. As this set of people start utilizing and earning from the platform, there will be an increase in the standards of living as compared to their previous state.

Undoubtedly, this will go a long way in the creation of employment opportunities for people that have the required skills. Even though some don't have, they will be eager to learn and attain the skills required so as to fit in and be able to execute one or more tasks on the platform.

How Fair & Instant Payment Will Impact Adaptability.

The method Effect Force has chosen to reward workers immediately after the completion of an HIT is actually a very good one, since it is based on the quality of task an individual is able to deliver within a certain period of time. Moreover, this payment method is instant, and since everyone will be rewarded accordingly, there would be a zeal or enthusiasm to carry out the task(s) with the maximum level of skills acquired and the best of work would also be delivered.

All these put together would eventually bring about adaptability on the platform.

Micro-tasking; A Panacea To Major Problems.

Going by the modus operandi of the Effect.IA platform, it would be seen it basically adopt the method of micro-tasking where large volume of job is splitted into small tasks that can be distributed over the Internet to many people, so as to be completed early. As to number of problems that can be solved by micro-tasking, there are no limitations. These include;

Accessibility, convenience and time constraints, and decision making; Microworks are easily accessible by anyone as long as he or she is using the appropriate device(s). Workers can work from anywhere in the world and at their convenient time and are able to decide the kind of job they want to do. There is no one to dictate for them. Therefore micro-tasking is able to solve the problem of accessibility, convenience, time constraints and decision making on the workers side.
Also, there is no need for a great acquisition of skills to participate in this kind of project.

On the requesters side; The employers are able to set up their Jobs online and be able to get various results from the workers. This help to bring his or her job to completion on time since the service offers a large workforce which completes tasks concurrently.

The fact that Effect Force can provide companies an outlet to completing/testing micro-tasks based project on a massive scale, make it a plus for the employers since any Job that is wrongly or not properly done will be fished out.
Say for instance, I set up a job in which workers are to upload various kinds of flowers and include their names without repetition. In a case whereby same flowers are uploaded with different names, it would be easily pointed out and rejected by the requester based on the service offered by the platform.

In this case, it helps to solve the problem of REPETITION.

Meet The Advisors and The Effect.AI Team

And a Glance at AI Token Sale Details

Summary & Conclusion

Looking at the various problems encountered in the development of AI, ranging from getting enough data, diverging tasks and high computational costs make it seems tiring to delve into this project. However, with the project brought up by the Effect Force and the solutions it tends to proffer, there will be thousands of people going into the development of AI, thereby creating employment opportunities and this will eventually turn the present world around.

For More Info & Resources

• Effect.AI Website
• Effect.AI Whitepaper
• Effect.AI Lightpaper
• Effect.AI Youtube
• Effect.AI Medium
• Effect.AI Facebook
• Effect.AI Twitter
• Effect.AI Telegram
• Effect.AI Review/Explanation Video)

This is my entry for @originalworks contest.
Click here to join the contest.

effect2018