Topic: A Decentralised platform for Translators! (Read 286 times)

Why don't you go with Bitcoin or Ethereum? I think that's more attractive to users because they can get instant liquidity, while there is no clear reason why your tokens will have values in the market.

Thank you for your comments, its important to note we didnt do a ICO we using our own funds. The token is needed thou and will remain as part of our project.

Why are we doing this ?

www.translateme.menu is a business founded in 2016, no relation to crypto no blockchain solutions.... think of this platform as to why we started our project.

Problems experienced with using google translation API for this business  Translateme.menu

1. Primary reason : costs are extremely high for serving translation over google network

2. API call is limited to 2000 characters - This created a issue with larger menus and a work around was achieved but not ideal

3. no support for languages out side of the top ten popular languages, the test revealed low quality score that was not even acceptable  for menus.

4. Privacy all docs submitted to API become property of google - Our platform could be extended to static printed docs but the privacy policy prevents this.

5. Usage policy prohibits white label solution

The team worked in 2018 to find solutions to these problems, leading to translateme.network as a overall solution to all.

Machine Translation is moving into a decentralized approach to training larger models on multiple machines, the future of better trained model is across thousands of machines. We pioneering this approach in translation and will give us an advantage if focused on. We will be releasing a mining algorithm that pays your for your machines resources, estimated BETA release in JUNE 2019

Review : https://www.openmined.org/ they are open source non profit moving in this direction with a proved distributed model that uses blockchain technology with no token or incentive to avoid the conflicts presented currently. the point is blockchain is used and a fundamental technology piece of there project.

Review : https://www.deepl.com/en/translator they compete with google directly and have successfully created an alternative solution thats profitable.  There approach was to scrape the internet of data used to build there models today. Capital and hardware and expertise needed but they have taken google on.

Review : telegram.translateme.chat it shows how translation can improve a major software solution like google, we built this to demonstrate users can contribute corrections and be rewarded with very little effort. This data is used on our platform and is a world first as a hybrid of AI and human input to create a truly unique algorithm that learns faster than traditional machine learning. It opens the doors to new data.

We taking the best of these examples and combining them.

The token is needed to move the value a company like www.translateme.menu would pay for these services to the contributors and is the last piece of the puzzle to bring this all together.

we are a small start up but we passionate about solving this original problems we faced with our own business in Mauritius, if enough people supported this project we  could collectively solve for this faster. Its the power of micro contributions on a large scale.


 

Thank you for your reply, take a look at this independent translation solution:  https://www.deepl.com/en/translator

Small company that has higher scores on there focused languages, the quality is important and we will continue to strive for better results. Commercially we wont need to achieve googles current waulity to be utilised on various meduims online, the easiert commercail gap is to partner with tourism industries on chat platforms, reviews, support and communication for basic booking information.

 
Decentralisng the network means we can rapidly grow with out capital. Our major advantage is these machines would help us resource data needed to imporve the output.

Our algorithm is unique, there is no company in the world that will incentive's contributions of human corrections of machine translation. PLease try our MVP it demonstrates this reward and contribution steps. www.telegram.translateme.chat
This data is unique to our network, google wont have access to this kind of data, they need more data to improve there mode. We will also make more effective translation in smaller languages and communities. Our network scales with out capital outlay learns from human input and rapidly increases quality compared to normal machine learning approaches. The network at our disposal will also scrape and collect existing data on the internet to achieve similiar results to DeepL.

Our own business in Mauritius www.translateme.menu is the reasons we started translateme.network our goal is to reduce the cost by 95 % and pay people directly.

Guys, stop questioning the relation of the project with crypto if the only application you know is to use it as a storage.
The main purpose is written in the OP, non-AI based translations which will be manually done by users that will be be paid in crypto (Only tokens which is unconvincing).
The reason must be: most projects and developers prefer human translations than GT or other AI.

However TranslateME, adding "decentralized" to the name of the project just because it's crypto-related IMO sounds unimaginative and dumb in the eyes of digital currency "tech-guys".
As the matter of fact, I've seen too many projects promoting them as "decentralized something" that were too hilarious that most doesn't even made sense.

Paying in crypto will not make the project decentralized.

Oh, I am quite surprised by this. I honestly thought that the platform would be one that had dedicated translators and would work alongside crowdfunding projects like stos/icos and translate their projects for a small fee. Sort of like a translators hub, but with each of them being verified and legitimate.

I don't think google translate is the best it can be right now, especially with the ai and technology we know they have. You'll be competing alongside one of the largest companies in the world, as startup that has little to no funds, I'm guessing?

Good luck, but I don't think just adding the blockchain to your project means it's able to compete with companies like Google, that's just impossible even for a lot of established businesses.

I agree, I don't see blockchain making any difference here. Probably the OP just wanted to use the technology behind, but it doesn't really makes sense at all. And who is going to pay to help improved the AI? It's hard to compete against, Google Translate with all the funding behind, I don't know if the project can really compete at high level, sorry to burst the bubble to the OP.

I tried DeepL, looks good but the language is limited. At least it doesn't use any blockchain to do its job, and that's fine and probably better imo.
We don't need blockchain for online translation.

If they are offering the services for free, why not? DeepL is a free translator. So if you need to pay something, then, that's another thing to think about. But if it will give you a better work than the free one, especially if you are working on technical documents that need better translations of your work, then someone might consider using that app. They can always get specific clients for this. They just need to expose their project to the right customers.

Regardless of the project of TranslateME. There exists a translation service which makes better translations than Google: DeepL .

What do we need blockchain for this? At what stage blockchain is used? To store data? Isn't it more efficient if we use IPFS or simply centralised storage?

Using blockchain for storing translations would be a waste.

Deepl's translator is one of the best there is for personal use.
In some cases, I can't even spot the difference from a translation made by humans.

Try it out for yourself, I use it a lot to translate foreign articles.

https://www.deepl.com/en/translator

I generally do not find any problem in google translate  but I guess it has few problems.
1. Sometime it translates words into word that mean the same but nobody uses them in the daily conversations.
2. Sometime it translates the word to its  literal meaning when it is not required.

Some human input can really improve the Google translation but why somebody should do double effort.

I think it would be correct: human translation will never be replaced by machine translation (I guess it's just your typo).
Google translater was stated by users above me. Well, to beat google in this situation, you can prioritize small and less knownable countries because in such countries google translator usually sucks and some of then don't even have speech. You can take it as an opportunity to develop in this aspect and push people in such countries to use your app instead of google.
Btw can't understand what does decentralization has to do here.

What is the role of decentralization here? These translators must be verified "centralization" so that their assistance to artificial intelligence systems is useful.

Also, the role of blockchain? Do you pay in NEO or TMN only? I checked your Youtube channel and understood that people are translating projects and paying them through a central site.


A real challenge awaits you, Google has developed its translator well.

Accurate. Plus, the unreleased projects by Google and other scientists are way ahead of whatever this company an achehve. Their training models are highly accurate and team of extremely smart people are improving them everyday. Google translate is just a beta version of what the general AI is and it’s not even as good as the paid translation services. It doesn’t matter what really you have in there, you will never be sponsored by the Fortune 500 companies to enhance your project.

Looks interesting, but my question is, can it really do something better than Google Translate ? I mean with all the money and compute power google has, I doubt that someone can train the AI better than them.

TrnalsateMe is a project which uses the power of AI, Blockchain and Human Input....and that's where translators come in.
 
We have already started on our platform to help translators and start building a community of translators that will help us build a strong network. Take a look at the video here : https://www.youtube.com/watch?v=kZU2ZmYgnDk

These community members will be great contributors to our work. Translators will be paid in NEO or TMN and earn extra NEO for helping improve the network's algorithm for machine translation. Machine translation will never be replaced by human translation but its needed in a lot of applications like our MVP we built www.telegram.translateme.chat for real time live translation machine translation is very important.

Stay up to date with the progression of this platform on our Twitter page, where we will soon be hosting a Treasure Hunt! https://twitter.com/translatemenet