TuringNet AI Platform AMA Summary-09/13/2018

Bella Wang
TuringNet
Published in
4 min readSep 13, 2018

Dear Valued TuringNet Fans!

We had a great pleasure hosting our first AMA (ASK ME ANYTHING) this week in our Telegram group. Below you can find a summary on questions asked and answers.

Please note that we will be hosting bi-weekly AMA, and will start collecting questions for the next AMA session shortly on Reddit. You may also submit any questions to the TuringNet team via this form.

Q1. stefano197409: How to use the computing power from miners crypto currency for learning models of artificial intelligence? Why Video Games they allow not only easier, faster and cheaper to conduct experiments and train algorithms, but also can become the basis of a new theory of intelligence this is really true in general how?How exactly does this happen now and will happen in the near future?

A: Dr. Wen: Good question. Let me answer the first part first. There may be many different ways to learn AI models through blockchain mining activities. However, we are focusing on our own and patented consensus algorithm — Learning BFT. It allows miners to update model parameters in every iterations and get reward if they calculate the best results.

For the second part, I am not an expert in Video Games AI, but a fan in video games too. What I can say is that Video Games are suitable for unsupervised learning, because they have well-defined rules and the games themselves are the actual simulators. With such simulators, we can train the AI models through zillions of experiments that lead the model towards the correct goal.

Q2. stefano197409: Currently, almost all computer vision is based on convolutional neural networks. Firstly, how do neurons convert input signals? Secondly, how neurons try to model how the brain processes information?

A: CNN is a very sophisticated machine learning model and answering the question needs a full class. In a nutshell, a neuron acts as a gate with a certain threshold. If the input signal strength is above the threshold, the neuron outputs 1, otherwise, the neuron outputs 0.

Similar to the brain, the neurons are connected in a way of many layers. A neuron in a layer can connect to multiple neurons in the upper layer, then the input signal takes a weight sum of the connected neurons’ outputs. After the processing of many layers, the final outputs become the prediction result of a CNN model.

Q3: stefano197409: Progress in artificial intelligence is taking place at a staggering pace. I have a question that always motivates me, and so much I want to know why more companies take only AI and automation so seriously?

A: Companies are always seeking for ways to improve their production efficiency. AI and automation play important roles in this trend. With AI, the company does not need many people to do the work.

In addition, the high speed and high precision of AI in some scenarios enable new opportunities, for example, online advertising can be benefit from personalization models to improve performance in the speed of millions of queries per second and among PB of data.

We also believe us humans are designed for something even smarter and more complicated. So AI will be our future assistant to free us from tedious tasks.

Q4: ntttairdrop: How does TuringNet stand out and firm compares to other AI decentralized platform projects in the current and future market? Any confirmed partnerships yet?

A: TuringNet’s competitive advantages over other AI decentralized platform projects are our true AI consensus, as well as the system architecture design and virtual machine built on top it. Such advantages create the new concept called decentralized AI models. Please refer to our whitepaper for more details.

Q5: Aljoinc: And how confidence and promising is the Community Mining Offering (CMO) as compare to the usual ICO? What will make sure its successful adoptions to early investors.

A: he team will neither pre-mine any TNET tokens nor do a presale to the community as ICO. Instead, all TNET tokens are mined through the community, while a portion of tokens are donated to the team and the investors.

The new CMO model creates a long-term binding between the community and the team and the investors. In this way, the tokens will be widely distributed among active community members, like Bitcoin, and be difficult to be controlled by any party. So the token price will truly reflect the project’s true value.

We will share more details on how you can participate from the very beginning, finish some fun mining tasks, and receive tokens in return.

Q6: Gloria Uchendu: How does one who is not interested in mining get token? Will you have your own exchange?

A: We will be offering various ways to for our users to get TNET tokens besides mining. Exchange is one of the options, while we are not yet considering building our own exchange. Other options include: 1) purchasing tokens from our company at market price to use AI services on the platform; 2) create your own AI models and get tokens paid by the clients that use your models.

About TuringNet — World’s First Open and Trustable Platform for AI

TuringNet is world’s first open and trustable Artificial Intelligence platform. The platform sets the future development and applications of AI free from centralized organizations, and reinvents how the public participates in scalable and collaborative model training and inferencing for various AI development needs throughout its entire lifecycle.

We aim to lower the barriers for all enterprises to adopt AI in their products and services, at the same time empowering a truly self-governing ecosystem to crowdsource the improvement of AI.

Where to find us:

Telegram Community:
https://t.me/turingnet

Twitter:
https://twitter.com/Turing_Net

Blog:
https://www.medium.com/turingnet

Github:
https://github.com/turingnet

Website:
www.TuringNet.com

--

--

Bella Wang
TuringNet

Business Strategist, Tech Futurist, Blockchain Enthusiast