SYPWAI is a decentralized AI platform for collaborative development. As part of the project, a business that needs analytics using neural networks meets with specialists who can implement the necessary tools. Besides, people here learn to work in this area while saving time and money.
Today we were able to interview Abdigani Dirie, Chief Operating Officer of SYPWAI, one of the most important people in the company
You said earlier that your team understands how to improve the accuracy of neural networks. Could you please tell me, how do you still manage it?
The creation of any AI project includes two different components. First, you have to get the data and tag it. Marking up information is a very simple process, but the work is monotonous and time-consuming. And since training a neural network requires learning about thousands of different types of data, his step is essential.
Secondly, you need to write a neural network itself. It is, in fact, a simulation of a small part of the human brain, which can then solve some problem, much like a human does. In the learning process, the neural network “is punished for” incorrect answers “and ” rewarded” for the correct ones.
This makes it easy to create any solution. The markers do their job, we check the execution accuracy, train the model with the received data, and create an API.
What is the main idea behind the SYPWAI project?
The main idea behind SYPWAI is that anyone can participate in neural network training. The process looks like an educational game for children. We look at the photo of the cat and answer who is this fluffy: a fox, a cat, or a dog. In this way, everyone can make money and benefit society at the same time. The main task of our artificial intelligence is to predict the best way to achieve this goal. This is achieved by enumerating a variety of possible development options.
What is more important in the SYPWAI ideology: social functions such as training neural networks or providing jobs for people with disabilities, etc.?
We have been working with artificial intelligence for a long time. In the process, we identified a big problem. Businesses often cannot allocate a separate unit under the data science, which will help in making decisions based on the neural network. Partitioning takes a huge amount of time. The global AI market is still in its infancy. There are no large companies or large-scale players here yet. We decided to create a structure, which will be useful to everyone. Therefore, it is impossible to choose a priority, everything is equally important here.
How does SYPWAI work in practice?
First, you need to complete a test task. It’s very simple. Therefore, our team will only allow those people who managed to get 99% correct answers to continue working. At this stage, tokens are issued symbolically, the price is hundredths of a percent. Then the person gets access to real work. He performs a certain part of it, which then is checked by other people involved in the project.
When most people’s answers match up, the data is sent to specialists for use in machine learning. The following principle works here: if seven out of ten people gave one answer, and three – another, we believe that the majority is already present. The correct answers have a positive effect on the rating and the level of earnings, the wrong ones, respectively, reduce the price of the user’s answer.
What about corporate clients? What questions do they come up with?
We receive a request from a company, a public organization, a scientific group that needs the help of our artificial intelligence. Besides, our organization itself creates a demand for solving the global problems of humankind. At the same time, corporate clients tell us about their main problems and goals.
For example, if a company develops prostheses, we will tell it how to make them better and cheaper. If a community organization is trying to determine how to reduce crime in a region, we train artificial intelligence to solve this problem.
In terms of customer surveys, we are preparing research studies and AI training questions for this company. Scientific Research – Artificial Intelligence helps you set the learning vector. Answers to questions from living people – help to identify the real needs of people so that products developed under the demand of people.
Tell us about the main algorithms for SYPWAI.
We study the characteristics of each client, collect maximum information about him and his field of activity. We clarify the goals, methods for solving the tasks, research of scientists in the right field, conduct surveys of real people, etc.
We encrypt all information using the XOR algorithm. Data transfers between devices and servers are encrypted using asymmetric methods. As a result, the possibility of information leakage is excluded.
It is important that when sending data, it is randomly mixed, which allows you to decentralize the exchange. The information is divided into small blocks to work with small, low-power devices. So, one block is sent to smartphones, and entire clusters are sent to large computers.
How much can a student who is engaged in markup from a smartphone earn in your project?
The average wage for marking is about $5 per hour. The student in your example earns between $300 and $400 per month. But here it is important to understand that almost any person can do this work.
In SYPWAI, payment is charged in tokens, which is a cryptocurrency. Given the latest trends, can this approach generate skepticism?
We are operating tokens, that’s right. But the cryptocurrency can be withdrawn to the bank card at any time or exchanged for some other Internet money. That is, you can earn your $5 per hour and immediately withdraw it from the system. There are no restrictions. But if the platform develops correctly, then the value of tokens will grow. Thus, each person decides for himself whether to withdraw his 5 dollars immediately or wait; perhaps in a couple of months, these tokens will cost 7 or 8 dollars. However, there are no guarantees. Here everyone decides for himself.