Partners and Investors & Token Sale

Right now Anryze is planning to raise token sale from $1.6 million to $6 million. This will help to speed up the growth of market, it is an official service provider and crowd funding initiator. Their plan is to create a special token on wave platform and name it RYZ, 100 million RYZ or 1.6 million$ to 321 million RYZ or 6 million $. The basic rate is 50 RYZ for 1 USD. Right now it is expected that 300 million RYZ will be offered with 20% discount for the buyers in first 2 days of token sale. Additional 6 million will be offered for bounty program and 15 million for rewarding employees. This makes total of 321,000,000 RYZ tokens without the discount and 381,000,000 RYZ in case the hard cap is collected in first two days of token scale. After an equivalent of 6 million$ creation of new tokens will stop and this will be sale expiration date. After the expiration of sale new tokens will not be created, all tokens represent the currency for service payment. In case the fund raising is less than planned money will be refunded. In case of any issue RYZ will be traded into crypto exchange in order to compensate the token owners and miners.

About the team :

The team consist of total 8 people, Anton Gera is the CEO and founder of the project whereas Mike Ezhov is CMO, Oleg Zaichuk is CTO, Max Kudymets is project manager, Oleg Fedosenko works as machine learning and development, Alex contributes to backup development and they have two PhDs from Moscow and Tokyo as their team members, name Sergey and Oleksii.

Decentralized Computing Network and Anryze:

Decentralized Computing Network Now the question is how decentralized computing network works?Anryze has decentralized worldwide speech recognition purpose. This as a result has a computing power exchange in free market. The expectation is that this will be cheapest and smartest decentralizing computing system. This is achieved due to self-learning intelligent agents. The network has three components, this includes computing power, a block chain wave for the payments and a node with distribution table. This system contains users as well as the miners, in short users are people that rents computing power for transcription purpose while miners, lease their power to get paid. User app just recognize the miner automatically via DHT and then transmit it to P2P link. After setting connection app connects the user and miner, later confirmation is received and payments is received by miner whereas, contract responsible for being the middle man get the commission. This whole procedure gathers business of 24h period and later, calculates the total amount for every wallet depending on the token. Simple transactions cannot use these calculations as they are quite expensive.

Description :

When the user upload file he receive the cost of transcription this is totally dependent on the duration of audio as well as the current price. When the user wants miner the program automatically connects to DHT and detect the suitable option according to the user. Ultimately, this sends request via P2P connection to miner and also receive the confirmation of readiness. Talking about the market or market size, problem is language barrier, you don’t understand what other is saying in case you want to record the message you still need to translate it later. This obviously means a lot of data is stored and recorded. Speech recognition is extremely helpful for business, as it makes text easy to analyze. In 2015 global voice recognition market size was estimated to be $51.09 billion, this is expected to grow with time at the rate of 11% per year. As the business is developing worldwide the adaption of voice enable application is also growing. Mobile phones and other smart gadgets are also using voice enabled consumer products. This is challenging the traditional voice technology engine growth that are contact centers. Healthcare sector also depends on the voice recognition solutions, this will help the sector to meet the EMR level and provide effective health care service. Voice recognition helps the doctors and soldiers to cross language barrier and accomplish better understanding in small time. The biggest investor in this market is military, the use this technology in order to increase the operational efficiency. The largest market is North America, Asian pacific countries are also catching up to the market shares.

Speech recognition platform with distributed computing network.

Language is one of the best form of communication that everyone uses. Where most of the time we learn the computer language most of the scientist and programmers are trying to come up with a process where computer can understand our language. This will help to create a more effective work bringing revolution in the field of programming and computer. Translation is never 100% accurate, for programmers this is quite challenging to develop something this accurate. Due to this problem there are other issues that are related to speech technology. They have come up with a process to help with the speech recognition. Anryze use a through speech recognition process in this process there are specific steps. The sound travels in the form of wave, this wave is split in words by the silence and this is then recognized word by word. Each word is matched with the already available set and once the combination is found word is recognized. As the number of words is more it takes long and they are trying to optimize that, but each word is usually 10 milliseconds long, the program extract 39 numbers that actually represent the speech, this is known as features vector. Number is generated via active investigation, in simple words it is derived from spectrum. Another important factor is model, it represent mathematical objects that combines the features of commonly spoken words. When we come up with an audio model it contains the combination of three states of senone, this is most portable feature vector. After the analysis it is found that there are some questions raised, how a model fit to the practice, is there a possibility of improvement for the internal model problem, how adaptive is it to the change and another important question is it matching the words itself. Matching is long process so optimization tricks are used to speed up the process. Once the words are matched, it extends them with time to formulate a framework that can predict the possible next matching variants.

Monetization :

Official website: https://anryze.com

Bitcointalk forum thread: https://bitcointalk.org/index.php?topic=1968881.0&utm_source=facebook&utm_medium=cpc&utm_campaign=conversion-audience&utm_content=NEW

White paper: https://tokensale.anryze.com/assets/file/Anryze%20Distributed%20Network_EN.pdf