Author Topic: [Worker Proposal] BitAsset PhD Research project - University of Colorado  (Read 76300 times)

0 Members and 1 Guest are viewing this topic.

Offline bitProfessor

There was no consensus within Cn-vote, but we started voting.

Offline biophil

  • Hero Member
  • *****
  • Posts: 880
  • Professor of Computer Science
    • View Profile
    • My Academic Website
  • BitShares: biophil

The main message of this paper will be that the specific choice of MCR and MSSR can have some strong and surprising effects on the risky behavior of BitAsset shorts (collateral holders), and that a small change in MCR can have a massive effect on the likelihood of a short position becoming undercollateralized. For example, I have one set of simulations that shows that changing the MCR from 1.6 to 1.5 can make undercollateralization 134 times more likely. Now, I'll point out that this doesn't mean I'm going to panic and immediately write a BSIP to argue for MCRs of 1.6; instead, it motivates me to look deeper into this phenomenon and start to understand how we can protect against it without harming liquidity and pegging. As many will be quick to point out, there is certainly a tradeoff involved here: reducing MCR from 1.6 to 1.5 may make the BitAsset less risky, but it will also make people more willing to provide collateral for it, which should improve price pegging. Understanding how to balance these tradeoffs wisely is a major focus of this project.


voted, very interested in how the simulation is done and what further conclusion is reached.

Thanks much for your support! I'm working on the paper today, and I'll provide a PDF of it to the community after it's been submitted to the conference in a few days.

Just to be clear and set your expectations appropriately, this first paper is analyzing a very simple model that isn't meant to capture all of the details of the bitshares system. The first step in a project like this is to analyze a simplified model, find where problems arise, and then use that to guide further questions. Basically, if we find warning signs looking at a simplified model, we expect that adding in the complexities of the real thing could make those problems even worse.
Support our research efforts to improve BitAsset price-pegging! Vote for worker 1.14.204 "201907-uccs-research-project."

Offline Digital Lucifer

  • Sr. Member
  • ****
  • Posts: 368
  • BitShares Maximalist & Venture Architect
    • View Profile
    • BitShares
  • BitShares: dls.cipher
  • GitHub: dls-cipher
TOP! I'll correct myself and real apology goes to Jerry (bitcrab). Thank you for finally hearing me out.

no need for apology, if you can I'd like to see you lobby more big proxies to support 1.5 as bitUSD MSSR. :)

It seems pretty impossible at the moment - as for atomic swaps, wallet UI and a lot of other very GOOD and NEEDED proposals, so I've decided to submit myself for proxy :)
Milos (DL) Preocanin
Owner and manager of bitshares.org
Move Institute, Non-profit organization
RN: 2098555000
Murska Sobota, Slovenia.

Offline bitcrab

  • Committee member
  • Hero Member
  • *
  • Posts: 1928
    • View Profile
  • BitShares: bitcrab
  • GitHub: bitcrab
TOP! I'll correct myself and real apology goes to Jerry (bitcrab). Thank you for finally hearing me out.

no need for apology, if you can I'd like to see you lobby more big proxies to support 1.5 as bitUSD MSSR. :)
Email:bitcrab@qq.com

Offline Digital Lucifer

  • Sr. Member
  • ****
  • Posts: 368
  • BitShares Maximalist & Venture Architect
    • View Profile
    • BitShares
  • BitShares: dls.cipher
  • GitHub: dls-cipher

The main message of this paper will be that the specific choice of MCR and MSSR can have some strong and surprising effects on the risky behavior of BitAsset shorts (collateral holders), and that a small change in MCR can have a massive effect on the likelihood of a short position becoming undercollateralized. For example, I have one set of simulations that shows that changing the MCR from 1.6 to 1.5 can make undercollateralization 134 times more likely. Now, I'll point out that this doesn't mean I'm going to panic and immediately write a BSIP to argue for MCRs of 1.6; instead, it motivates me to look deeper into this phenomenon and start to understand how we can protect against it without harming liquidity and pegging. As many will be quick to point out, there is certainly a tradeoff involved here: reducing MCR from 1.6 to 1.5 may make the BitAsset less risky, but it will also make people more willing to provide collateral for it, which should improve price pegging. Understanding how to balance these tradeoffs wisely is a major focus of this project.


voted, very interested in how the simulation is done and what further conclusion is reached.

TOP! I'll correct myself and real apology goes to Jerry (bitcrab). Thank you for finally hearing me out.
Milos (DL) Preocanin
Owner and manager of bitshares.org
Move Institute, Non-profit organization
RN: 2098555000
Murska Sobota, Slovenia.

Offline bitcrab

  • Committee member
  • Hero Member
  • *
  • Posts: 1928
    • View Profile
  • BitShares: bitcrab
  • GitHub: bitcrab

The main message of this paper will be that the specific choice of MCR and MSSR can have some strong and surprising effects on the risky behavior of BitAsset shorts (collateral holders), and that a small change in MCR can have a massive effect on the likelihood of a short position becoming undercollateralized. For example, I have one set of simulations that shows that changing the MCR from 1.6 to 1.5 can make undercollateralization 134 times more likely. Now, I'll point out that this doesn't mean I'm going to panic and immediately write a BSIP to argue for MCRs of 1.6; instead, it motivates me to look deeper into this phenomenon and start to understand how we can protect against it without harming liquidity and pegging. As many will be quick to point out, there is certainly a tradeoff involved here: reducing MCR from 1.6 to 1.5 may make the BitAsset less risky, but it will also make people more willing to provide collateral for it, which should improve price pegging. Understanding how to balance these tradeoffs wisely is a major focus of this project.


voted, very interested in how the simulation is done and what further conclusion is reached.
Email:bitcrab@qq.com

Offline Digital Lucifer

  • Sr. Member
  • ****
  • Posts: 368
  • BitShares Maximalist & Venture Architect
    • View Profile
    • BitShares
  • BitShares: dls.cipher
  • GitHub: dls-cipher
I'm well under way generating preliminary results for the potential Decentralized 2019 paper. If all goes well, this may be the first publication resulting from this project, and will pose several of the important questions that we will focus on.

The main message of this paper will be that the specific choice of MCR and MSSR can have some strong and surprising effects on the risky behavior of BitAsset shorts (collateral holders), and that a small change in MCR can have a massive effect on the likelihood of a short position becoming undercollateralized. For example, I have one set of simulations that shows that changing the MCR from 1.6 to 1.5 can make undercollateralization 134 times more likely. Now, I'll point out that this doesn't mean I'm going to panic and immediately write a BSIP to argue for MCRs of 1.6; instead, it motivates me to look deeper into this phenomenon and start to understand how we can protect against it without harming liquidity and pegging. As many will be quick to point out, there is certainly a tradeoff involved here: reducing MCR from 1.6 to 1.5 may make the BitAsset less risky, but it will also make people more willing to provide collateral for it, which should improve price pegging. Understanding how to balance these tradeoffs wisely is a major focus of this project.

To support our work, please speak with your proxy and ask them to vote for worker 1.14.204, "201907-uccs-research-project."

Many thanks to the proxies who are currently voting for our proposal:
  • evangelist-of-bts
  • openledger
  • xeroc
  • beos
  • clockwork
  • fractalnode
  • abit
  • bitshareseurope
  • bitspark-delegate

https://bitsharestalk.org/index.php?topic=28670 - re MCR 1.5 bitUSD

You should also buzz bitcrab for vote, how privately that publicly, he is a bit ignorant towards any other worker than his own, but in this case - pretty much aligned opinions between two of you.

Chee®s
Milos (DL) Preocanin
Owner and manager of bitshares.org
Move Institute, Non-profit organization
RN: 2098555000
Murska Sobota, Slovenia.

Offline biophil

  • Hero Member
  • *****
  • Posts: 880
  • Professor of Computer Science
    • View Profile
    • My Academic Website
  • BitShares: biophil
I'm well under way generating preliminary results for the potential Decentralized 2019 paper. If all goes well, this may be the first publication resulting from this project, and will pose several of the important questions that we will focus on.

The main message of this paper will be that the specific choice of MCR and MSSR can have some strong and surprising effects on the risky behavior of BitAsset shorts (collateral holders), and that a small change in MCR can have a massive effect on the likelihood of a short position becoming undercollateralized. For example, I have one set of simulations that shows that changing the MCR from 1.6 to 1.5 can make undercollateralization 134 times more likely. Now, I'll point out that this doesn't mean I'm going to panic and immediately write a BSIP to argue for MCRs of 1.6; instead, it motivates me to look deeper into this phenomenon and start to understand how we can protect against it without harming liquidity and pegging. As many will be quick to point out, there is certainly a tradeoff involved here: reducing MCR from 1.6 to 1.5 may make the BitAsset less risky, but it will also make people more willing to provide collateral for it, which should improve price pegging. Understanding how to balance these tradeoffs wisely is a major focus of this project.

To support our work, please speak with your proxy and ask them to vote for worker 1.14.204, "201907-uccs-research-project."

Many thanks to the proxies who are currently voting for our proposal:
  • evangelist-of-bts
  • openledger
  • xeroc
  • bitcrab
  • beos
  • clockwork
  • fractalnode
  • abit
  • bitshareseurope
  • bitspark-delegate
  • dls.cipher

Edit: added a couple more witnesses 7/29/2019
« Last Edit: July 29, 2019, 05:44:05 pm by biophil »
Support our research efforts to improve BitAsset price-pegging! Vote for worker 1.14.204 "201907-uccs-research-project."

Offline biophil

  • Hero Member
  • *****
  • Posts: 880
  • Professor of Computer Science
    • View Profile
    • My Academic Website
  • BitShares: biophil
A quick update:

  • Pending funding, my goal is to write a short paper to submit to the Decentralized 2019 conference in Athens. There is a separate worker which seeks to have BitShares sponsor the Decentralized 2019 conference, and I thought it would be fantastic if BitShares could also showcase the UCCS research project there as well. If all goes well, I'm going to scramble to submit a paper to this conference's academic track which will lay out the project's main questions and perhaps start to look at the shape of the answers.
  • The Research Project's worker period has started today, and it still needs more votes.
Support our research efforts to improve BitAsset price-pegging! Vote for worker 1.14.204 "201907-uccs-research-project."

Offline biophil

  • Hero Member
  • *****
  • Posts: 880
  • Professor of Computer Science
    • View Profile
    • My Academic Website
  • BitShares: biophil
Thanks for your reply, BTSMoon. Let me offer a few answers to your concerns.

I appreciate your effort to conduct BitAsset PhD Research project but I disagree Bitshares need to funds the PhD student at the University of Colorado for 2 years.

Why I Disagree this worker proposal:
1. Cryptocurrency world change by day, not by year. What work today, may not working after 1 year.

I agree; to combat this, you need generalizable solutions. The project's goal is to create a framework which allows us to study BitAsset design comprehensively so that we can have plans in place to deal with changing market conditions. I'm curious -- what do you recommend instead of making contingency plans?

Quote
2. This worker proposal measured by time rather than work which mean just need to wait time pass to get $$$.

First, as the proposal mentions, the money doesn't come to me; it comes to my University. The University of Colorado has many safeguards in place to prevent me from running away with the money. That is, it's literally not possible for me to just wait till the worker expires and then disappear with the money.

Second, I technically could take the money and then leave it in an account and not spend it. But that would be a stupid thing for me to do, because this is all happening in public: I'm not anonymous (as a professor, I'm quite the opposite), and I have my entire future career to spoil if I mis-spend this money. If I take your money and then don't produce anything meaningful, that would make it extremely difficult for me to ever do it again.

I can't do good research unless I can find good funding sources; if I waste a source's money, that will severely hurt my ability to convince other funding sources to fund more good research.

Quote
3. Workers proposal for such long period may voted out half way due to high competition. In that case, what happen to this worker proposal? Most likely is "We can't continue because no more $$$" and Bitshares get nothing at the end.

The work is measured in time, but the worker provides the money all at the beginning. At current market prices, the $160,000 budget could be released in as little as 2 months; the project will continue for 2 years after this with no further funding. I know this is unusual for a worker proposal, but it's not possible to attach the PhD student to the project unless I have the ability to make a long-term funding commitment to him.

Quote
4. Your BTS accounts biophil and browns never create any BitAsset, not even buying and selling BitAsset. Basically you are good in talking but not in action and only know BitAsset others people words. I don't foresee you can help us to solve problem. What I foresee is just academic talk.

Let me ease your mind: I have been active on the BitShares blockchain(s) literally since day 1. Consider these points:
  • biophil is account number 1361; this means I founded it on the BitShares X chain (prior to the launch of the current BitShares chain), before abit (12376), baozi (2204), fav (21106), cass (5178), etc.
  • "biophil" is not my main account. Take a look at account zebulon: http://cryptofresh.com/u/zebulon. To prove that I control zebulon, this morning I added biophil as one of the active permissions. Zebulon is a lifetime member account; I upgraded to a lifetime membership in block 8886: https://bitshares-explorer.io/#/blocks/8886, this is less than 8 hours after BitShares 2.0 main chain launched.
  • If you're worried that I have no experience with the BitAsset mechanisms, feel free to explore zebulon's history to see that I've been quite active in the markets.
  • It's true that professors have a reputation for being good at talk and not at action, but I'm good at both; my activity on Steem gives some recent blockchain-focused examples: I created the STEEM sub-token POCKET (https://steemit.com/pocket/@biophil/pocket-announcement) as well as a simple AI-based curation bot the "gentlebot" (https://steemit.com/introduceyourself/@gentlebot/hello-steemit-i-m-the-gentlebot).
Quote
5. Bitshares lack of budget 8).

Think of this in amortized terms: $160,000 over two years is $6666 per month, less than half the daily worker budget at today's prices. Essentially, the proposal is asking for less than a 60th of the BitShares budget to support a full-time student and 2 months per year of my time to study low-volatility assets exclusively.
Support our research efforts to improve BitAsset price-pegging! Vote for worker 1.14.204 "201907-uccs-research-project."

Offline iamredbar

4. Your BTS accounts biophil and browns never create any BitAsset, not even buying and selling BitAsset. Basically you are good in talking but not in action and only know BitAsset others people words. I don't foresee you can help us to solve problem. What I foresee is just academic talk.
5. Bitshares lack of budget 8).

My two cents on these two things:
4) You don’t need to use or create BitAssets to see from the outside how they operate. This is an ignorant statement. I could then say that you haven’t ever made a worker proposal, so how would you know how they work?
5) BitShares has the same budget, it depends on what we prioritize. PHD study and paper releases on a high level is a good move IMO.

Offline BTSMoon

  • Full Member
  • ***
  • Posts: 91
    • View Profile
I appreciate your effort to conduct BitAsset PhD Research project but I disagree Bitshares need to funds the PhD student at the University of Colorado for 2 years.

Why I Disagree this worker proposal:
1. Cryptocurrency world change by day, not by year. What work today, may not working after 1 year.
2. This worker proposal measured by time rather than work which mean just need to wait time pass to get $$$.
3. Workers proposal for such long period may voted out half way due to high competition. In that case, what happen to this worker proposal? Most likely is "We can't continue because no more $$$" and Bitshares get nothing at the end.
4. Your BTS accounts biophil and browns never create any BitAsset, not even buying and selling BitAsset. Basically you are good in talking but not in action and only know BitAsset others people words. I don't foresee you can help us to solve problem. What I foresee is just academic talk.
5. Bitshares lack of budget 8).
 

Offline biophil

  • Hero Member
  • *****
  • Posts: 880
  • Professor of Computer Science
    • View Profile
    • My Academic Website
  • BitShares: biophil
Greetings community! Two months ago, I (Dr. Philip N. Brown, Assistant Professor of Computer Science at the University of Colorado Colorado Springs) made the announcement here that I was embarking on a unique research project to study bitAssets. You can read my original introduction post and its discussion here: https://bitsharestalk.org/index.php?topic=28340.0

Today, I give you a worker proposal to raise funds for the project and ensure its success; many thanks to @xeroc and the BitShares Blockchain Foundation for their invaluable assistance in putting forward this proposal. You can find the complete proposal below and at https://www.bitshares.foundation/workers/2019-07-uccs-research-project; but here is the TL;DR version:

  • This worker fully funds a PhD student at the University of Colorado for 2 years; under my supervision, this student will study the fundamental design tradeoffs surrounding price-stabilized crypto-tokens with a particular focus on bitAssets in BitShares.
  • The student and I will present reports and recommendations to the BitShares community as well as author publications which we will submit to high-profile academic conferences and journals for peer-review.
    • Whenever we give a talk for this project at a conference, we'll acknowledge the BitShares Blockchain as the research's funding source.
    • Whenever we publish a paper resulting from the funds in this worker, we'll include the text "This work was supported by BitShares worker proposal 1.14.204."
  • The BitShares Blockchain Foundation will act as escrow, and the University of Colorado will be the custodian of the funds for the duration of the project (read: I can't just cut and run with the cash).

Please pepper me with your questions, and I will make myself available to chat via Skype (philskypetalk) or any other chat service if that would be helpful. I'll also update this post with something like a FAQ as I start to see which questions are the most frequent. What follows is the complete proposal as published at https://www.bitshares.foundation/workers/2019-07-uccs-research-project:

Name/bfid: 201907-uccs-research-project
Category: Research
Team: Dr. Philip N. Brown, Assistant Professor of Computer Science
Total: $160,000
Duration: 4 months
Worker ID: 1.14.204
Discussions: https://bitsharestalk.org/index.php?topic=28340.0
Accounting & Reporting: https://workers.bitshares.foundation/201907-uccs-research-project

University Research Project for BitAssets
This worker proposal, if approved and funded, will support a gift to the University of Colorado Colorado Springs (UCCS) to fund a 2-year bitasset-specific research project by a PhD student under the direction of Dr. Philip Brown. Please scroll down for more details.

Yearly Budget
By far the most critical budget items are the stipend and tuition payments for the PhD student, who will be committing the bulk of the time to the project. In addition to the student’s budget items, the project’s impact is significantly enhanced by including some time for Dr. Brown, to ensure that he can adequately direct the research. Time for faculty can be included as salary in the summer months or as a “course buy-out” during the academic year which reduces his teaching load to allow more time for research. The following budget includes a summer month as well as a teaching reduction. Finally, an annual $10,000 travel budget is included; both Dr. Brown and the PhD student will travel to universities, conferences, and workshops to promote the work, disseminate research findings and develop future collaborative relationships with other researchers. UCCS applies indirect costs of 10% to each budget line item below to cover facilities maintenance, support staff salaries and related grant management expenses; each of the following budget items includes these 10% indirect costs.

A sample annual budget is as follows:
  • PhD Student annual stipend: $26,736
  • PhD student annual tuition (6 credits per semester): $10,211.11
  • Dr. Brown month of summer salary: $14,354.44
  • Dr. Brown 1-course teaching load reduction: $12,919.00
  • Travel budget: $11,111.11
  • Total annual budget: $75,331.66
If the project is funded at higher levels, the additional money will be used to extend the student’s funding for up to a third year. It is important to stress that this is likely to significantly increase the ultimate impact of the project, as the productivity of PhD students tends to increase over time.

Accountability and Transparency
Once the funds from this worker proposal are disbursed to UCCS, the University of Colorado will be the custodian of the funds as required for all funded university research projects. Dr. Brown will have signature authority over the funds, but all expenditures will be vetted through the usual accounting procedures of the University of Colorado, providing a strong layer of protection against embezzlement and other malfeasance by Dr. Brown.

Budget Restrictions
Worker funds are released relatively slowly, so an extremely aggressive timeline may not be possible. However, the PhD student cannot be attached to the project unless a minimum of two years’ funding is allocated to him up front.

To reduce costs, this worker proposal may pay out BTS to the BBF (the escrow) directly to reduce trading fees and spreads that would otherwise increase total costs to cover the commitment. Payments to the University of Colorado will remain transparent to the Community and the voters.

Executive summary
Dr. Philip Brown wrote an executive summary that contains the full details of the PhD student to work on.

Read full executive summary: http://cs.uccs.edu/~pbrown2/files/BTS_executive_summary.pdf

Key Research Thrusts:
  • Recommend concrete improvements to the BitShares MPA system.
  • Develop a “science of collateralized tokens,” investigating the key design tradeoffs in the space.
  • Can the balance between short-term and long-term stability be tuned?
  • Is there a fundamental tradeoff in collateralized tokens between manipulability and short-term stability?
  • How can we assess risk in collateralized token systems, taking into account the incentives seen by all parties?

Additional project activities:
  • Travel to universities and academic/blockchain conferences to present our research.
  • Participation and outreach in blockchain community events in the greater Colorado Springs area.
  • Sharing ongoing research findings with the BitShares Core Team and contributing insights to relevant design discussions.

Principal Investigator Bio
Dr. Philip N. Brown, an Assistant Professor at the University of Colorado Colorado Springs, will direct the research project. Dr. Brown brings uniquely multidisciplinary expertise to the project, as his research combines concepts from automation and controls, game theory and incentive design, and integrated engineered and engineered systems. Dr. Brown has been a BitShares community member since early 2014, and participated in bitsharestalk.org discussions (under the pseudonym “biophil”) regarding the design of the first BitAsset pegging mechanisms. Since then, he has launched an interdisciplinary research program in the department of Computer Science at the University of Colorado: http://cs.uccs.edu/~pbrown2/.

Summary
The proposed research gift funds an exciting and ambitious undertaking to result in both technical results relevant to the functioning of the BitShares Blockchain as well as basic science which will advance knowledge and lead to visibility and lasting impact in the broader academic community.

It is our hope that in its success, this project will accomplish the following three unique things:
  • Bring new tools and theory to bear on the practical challenges faced by the BitShares Blockchain,
  • Spread awareness of the BitShares ecosystem to a previously unreached audience, and
  • Showcase the BitShares community’s ability and commitment to fund higher learning projects which confer benefits beyond the insular setting of the BitShares Blockchain itself.
Support our research efforts to improve BitAsset price-pegging! Vote for worker 1.14.204 "201907-uccs-research-project."