Garbage in - Garbage out - the only reason why Google or Amazon might be best at developing constantly self-improving neural networks for things like speech recognition is because they have an insane amount of data. How are you going to bootstrap that problem? Because your neural networks will be extremely weak in the beginning and the output is probably useless for quite some time. Am I wrong?
You are not quite right. Google and Amazon might have large volume of data because they have been in the business for a long time, they too started in a day. We our technology running on the blockchain platform that offers value for both miners and users, we compete favorably with them.
In using Anryze, the removal of central controls would mitigate most traditional data failures and provide the highest accuracy of a speech transcription due to neural networks education through distributed system.
What wasn't right from what I said? I just asked you how you are going to get enough people on board so that you can really gather sufficient data and train your neural networks well? It's a tough problem to solve, and ultimately you are still a niche in the blockchain space. Not saying you don't have a chance, but this will need a looooot of time I assume.
The amount of data will come from call centers that generate fairly large data flows, for example an average call center (~200 people), provides 1.5 million minutes of calls per month.
More information you can get from the WP
