Get Paid to Train AI with Dbrain

Now, you can teach robots to think like we do. Don’t believe me? Take a quick look at any of the sources below about the Dbrain project, and you’ll begin to see what I mean.

Even so, you might find some of your questions unanswered.

In this short post, I hope that this becomes clearer for you and I also hope that it just might pique your interest in getting involved in the industry, even if you don’t have a highly mathematical background.

What is Dbrain and what drives Dbrain?

Dbrain runs on the premise that the vast majority of an AI system is data and the success of such a system is completely dependent on having and acting on the “right” data. The center of Dbrain’s offering is what appears to be an easy to use interface with which human users categorize data, which is then shared with AI systems that participate in the Dbrain network.

The Sharing of Data with Dbrain

Most important to understand Dbrain is understanding how its chief product works.

First and foremost, it’s a Blockchain-based product. When you do tasks on the network, you are paid in Dbrain coins. Once a payment goes through, Dbrain also states that you don’t have to wait at all to withdraw your coins. If this, in fact, turns out to be true when the network is launched, then one can’t help wondering: what’s the primary incentive to get on and stay on the network?

Secondly, it does not appear that Dbrain actually plans to directly connect Artificial Intelligence networks to their system. They actually plan to funnel the categorized or labeled data from labelers to data scientists or anyone on AI teams. If this is done for the lifespan of the network, then this effectively means that there will be no direct AI involvement in it. If AIs truly aren’t involved, then a key question comes to light, which we’ll touch on below. Before getting there, it’s important to mention two more key features of the Dbrain network that cater to a full understanding of what it does and why it does it.

Overall, Dbrain plans to include four distinct groups of users. The process of using the platform will begin with owners of data, which could conceivably be AI teams, wo upload data to be categorized.

At this point, labelers or individual human users will begin to work on sorting the data. Once they are done, Dbrain’s consensus protocol validates the results that are obtained.

Dbrain’s SPOCK Protocol

This is where it all gets even more interesting. Dbrain is yet another Blockchain project that wants to veer away from traditional Proof-of-Work and Proof-of-Stake consensus protocols. Dbrain’s consensus protocol with regards to the data that is submitted as categorized by users, is called the SPOCK protocol. Precisely, SPOCK is an acronym for “Subjective Proof of Crowdwork,” and Dbrain claims that it validates the users’ output data automatically and it helps to facilitate on-time payments to users, as well. SPOCK can and will reject what it terms “incorrect” offerings by labelers and according to Dbrain, this means that the labelers in question will be penalized with low ratings on the network and no pay.

Given all of this, it appears that SPOCK could be Proof-of-Work based, though to date, there doesn’t seem to be any specific mention of computing power or nodes needed to run this protocol.

Upon a closer look, it becomes evident that validators are users that weren’t mentioned in Dbrain’s earlier description of how their network is meant to function. Validators are participating nodes that approve submitted data with the help of the SPOCK protocol. Therefore, SPOCK appears to be Proof-of-Work, and we also end up with five expected groups of users that cater to the functioning of the Dbrain network.

The Importance of the PICARD Protocol

None of what’s been mentioned already relates to keeping the Dbrain network secure. That’s where the PICARD Protocol comes in. Here, PICARD stands for “Protocol for Indirected Controlled Access to Repository Data.” All in all, this algorithm secures all of the data on the network. With his in mind, PICARD needs to be iron-clad in its efficiency, given that a data classification network like Dbrain could lose all credibility after one major hack. The lifeblood of any profitable work with data is the fact that all networks involved, consistently stay secure.

If they don’t, then these networks lose the trust of its users.

PICARD not only secures the network but in doing this, it also allows AI teams that are signed on to Dbrain to work with the data that they receive there, online and securely. Given this as well as the mention of enabling community projects on the network, there are also implications towards Dbrain becoming much more noteworthy, if all goes well.

What’s left unsaid? Future of Dbrain

Future pieces on the specific features that seem to imply that Dbrain will become the chief AI platform of the future, might be even more illuminating. Even though it’s been stated here that Dbrain doesn’t directly involve AI in their network, that’s by design, at least for now. All signs seem to point to Dbrain being the platform that helps the everyman feel comfortable with getting involved with AI projects. At the same time, all signs also seem to point to Dbrain being the platform that reduces the monotonous part of a data scientist’s job, significantly. Without the need for cycling through mounds of data to identify it for “Deep Learning training sets,” their time can be freed up for more high-level activities like monitoring and altering algorithms as AI systems run.

All in all, here’s to hoping that Dbrain achieves all of this and more and even expands the human involvement on their platform, so we can eradicate the irrational fear of machine-intelligence growth.