Tuesday, May 29, 2018

Some blockchain applications and Reviewcoin

Conjecture 34: Forall x, we have "x on blockchain".
I don't have a proof but I have seen dental on blockchain, kiwis on blockchain, shoes on blockchain, etc.

Apart from the many silly x-on-blockchain attempts, I have heard some serious and promising applications. The Brave browser basic attention token seems to be a serious effort and can help redefine the ad-economy on the browsers. I also recently heard about Abra digital currency bankteller, which is another solid company and application.

To continue, Jackson Palmer said these are his favorite decentralized technology projects: WebTorrentDat / BeakerMastodon, and Scuttlebutt. Finally, there are well circulated requests for apps, which I think can make some uptick in the blockchain applications game.

My blockchain app suggestions

I didn't want to be left out of the action. Ideas are free. Here are some things that occurred to me.

Medical school students can do an ICO and fund their schooling

This will be a utility token, where the owner can get 30 minutes visitation/consulting worth per token. Of course as the student becomes a more experienced and skilled doctor, the token will be more valuable, because 30 minutes of the doctor is now more valuable. Of course with the token comes a token economy, where you can buy/sell tokens for doctors. 

The same idea can apply to some other graduate schools as well, including electrical engineering, computer science, etc. Maybe some labs can raise funding for students/projects using this model.

Being a philanthropist never was so profitable/greedy. And being greedy was never more philanthropic.

But come to think of it, I don't like this. Would this lead to maintaining a stock ticker for people in your life? (Hmm, his grades this semester are not that good, time to short his tokens.) People into stock trading all day may be OK with this, but definitely not for me.

Review coin

This idea occurred to me recently: We should build a blockchain solution for conference reviewer crediting system. If you don’t help with reviewing, you don’t get to have your papers reviewed.

The biggest advantage of doing this in blockchain rather than using a traditional centralized databases would be that now this will belong to everyone in the academic community rather than a specific organization. IEEE versus ACM is a thing. There are many other organizations. Even some universities may not want to submit to a system if it is maintained by a "rival" institution. (Though I must say the arxiv.org owned by Cornell gained adoption from all/most institutions as a repository of electronic preprints, so it looks like it can be done.)

With blockchains technology it may be possible to get the incentive model right as well, since similar problems have been studied in blockchains for cryptocurrencies.

There are many challenges here though. We like to keep the reviewers anonymous. We also want to keep the power of the anonymous reviewers in check and avoid bad/empty/trivial reviews and enforce a certain quality-level in the reviews.

To keep the review quality in check we can require multisig on a review. The extra endorsements may come from PC chairs or some assigned reviewer-reviewers to keep to a more decentralized system.

The participants (authors, reviewers, reviewer-reviewers) can be anonymous. The reviews do not need to go into blockchain/dag, only hash of the reviews can go in there for verification while maintaining some privacy. The participants will have private key and public keys, so it is not possible to initiate a transaction on behalf of another as in Bitcoin. On the other hand, we are still worried about double-spending and attacks on the blockchain.

To fend off attacks, the Reviewcoin blockchain could be powered with PoW, PoS, or leaderless-gossip-based (as in the Avalanche paper). This is a crazy thought but would it at all be possible to make this blockchain reviewer-work powered? There are some works that look at human-work puzzles and proof-of-useful work for blockchains. I speculate how a review-work powered method could be adopted for this problem in MAD questions below.

Reviewcoin can enable a better compensated yet more open model for reviewing papers. The model can also be applied to reviewing whitepapers, component designs, or even code-reviewing. The party that needs to get something reviewed should hodl some Reviewcoin, which could be earned by either doing meaningful reviews, or helping with Reviewcoin blockchain infrastructure running strong, or via purchasing from Reviewcoin owners.

By offering more Reviewcoins it can be possible to recruit more skilled/capable/experience reviewers for the work. Offering more coins should not get you more favorable reviews though, and the reviewer-reviewers should be the enforcer for fairness appointed by the system.

MAD questions

1. Is linking reviews with compensation opening a can of worms?
Probably... But the existing system is not great either. Academics volunteer for technical program committees of prestigious conferences (1) to help, (2) to get some name recognition as your name gets listed on conference website, and (3) to keep current with new research in the field. This is a good-will based model; these reviewers would consult to outside starting from $200 an hour yet they spend tens of hours for free to review for a conference. Inevitably this good-will based approach is starting to face many problems. Freeloading is a problem, as we are seeing some conferences getting 100s and even 1000s of paper submissions. Incentives is becoming a problem: for journals and weak conferences, it is hard to find reviewers, because the name recognition part is not that strong. Finally, experts are overstretched with volunteering.

So maybe we should consider opening a can of worms, if it can help us build a more sustainable system going forward.

2. Human proof-of-work for Reviewcoin
To certify that a review is valid and not random junk to get some Reviewcoins, the review can be checked and signed by program committee chairs if this is for a conference. We can have authenticated academics serving for PC chairs, so their signature certifies the review is correct and the reviewer gets some Reviewcoin for the service. To extend to non-conference unorganized reviews,  the review certification can be done through a multisig from a randomly selected set of participants. If the Byzantine nodes (Sybil/bot participants) are in the minority, they will not be able to issue false certification of reviews. It may be possible to use a proof-of-stake of participants to limit the number of Sybil nodes. Also the Reviewcoin stake/credibility of participants can be used in a weighted manner for reliability.

3. Where did May go?
Time is an illusion. End-of-semester doubly so. (With apologies to Douglas Adams.)

No comments:

Two-phase commit and beyond

In this post, we model and explore the two-phase commit protocol using TLA+. The two-phase commit protocol is practical and is used in man...