Milestone Release 1 |
$2,400 USD | Transfer Complete | 16 May 2024 |
Milestone Release 2 |
$10,970 USD | Transfer Complete | 14 Jun 2024 |
Milestone Release 3 |
$11,080 USD | Transfer Complete | 19 Jul 2024 |
Milestone Release 4 |
$11,660 USD | Transfer Complete | 12 Sep 2024 |
Milestone Release 5 |
$13,890 USD | Transfer Complete | 10 Oct 2024 |
We simulated how people vote on several voting systems, including the one currently used by sNET, 1c1v and other promising alternatives. We also considered how strategic and extreme voting impact the results. This research is critical to designing even better voting systems that optimize the aggregate voter's welfare and provide better results to the community.
SingularityNET is seeking to advance its high-quality decision process for its decentralized Deep Funding program. In support of this aim, Photrek proposes to develop an open-source simulator which models how the complex network of community sentiments impacts voting outcomes. The project will both lay a foundation for voting simulation capability and demonstrate an initial result. This first investigation will evaluate the claim of Plural Voting that applying a quadratic cost to the strengths of opinion assures a) truthful voting, and b) mitigation of controlling factions. A roadmap of future capabilities including simulation of collusion detection and reputation scores will be provided.
New reviews and ratings are disabled for Awarded Projects
Overall
4.0
4.0
5.0
Desirability
The proposal addresses a critical need within SingularityNet's decentralized Deep Funding program by developing an open-source simulator to model the complex network of community sentiments and its impact on voting outcomes. The project's focus on evaluating different voting methods, such as plural voting and no-cost grading, to maximize the quality of community decisions is highly desirable. Additionally, the incorporation of Monte Carlo simulation techniques and statistical analysis adds rigor to the evaluation process. The proposal outlines clear objectives and milestones, indicating a strong commitment to advancing SingularityNet's governance capabilities. Further clarification on the specific methodologies for modeling community sentiments and implementing voting simulations would enhance the desirability of the project.
Viability
The project's viability for SingularityNet AI is evident in its potential to enhance the decision-making process within the Deep Funding program. By developing a simulation tool to evaluate different voting methods and their impact on community decisions, the project aligns with SingularityNet's mission of democratizing AI. The proposed milestones and deliverables align well with the project's objectives, demonstrating a practical approach to implementation. However, the proposal could provide more insights into the scalability of the simulation tool and its long-term impact on SingularityNet's AI ecosystem.
Feasibilty
The proposal demonstrates feasibility through a well-structured plan and budget allocation across five milestones. Milestones such as literature review, prototyping of Monte Carlo simulations, and analysis of voting dynamics provide a clear roadmap for project execution. The team's expertise in statistical modeling and simulation, as evidenced by their previous contracts with IOTA and Cardano Catalyst, adds credibility to their ability to carry out the project successfully. However, the feasibility could be further enhanced by providing more details on the technical specifications of the simulation tool and the integration process with SingularityNet's platform.
Usefulness
The proposal offers significant usefulness to SingularityNet AI by addressing critical issues related to decision-making and governance within the Deep Funding program. The project's focus on evaluating different voting methods, incorporating community sentiments, and providing insights into voting dynamics is invaluable for improving the quality and fairness of community decisions. The open-source nature of the simulator also allows for transparency and collaboration within the SingularityNet community. The proposal demonstrates a clear understanding of SingularityNet's objectives and presents a valuable initiative to support its growth and development.
Overall
5.0
5.0
5.0
The proposal targets a highly undervalued research area that is crucial for the survival of any intelligent self-organizing system including Deep Funding and SingularityNET. For this ecosystem, it is not a must-have project quite yet but it will be at some point in the near future and it is definitely not too early to start working on it.
These pretty straightforward simulations won’t yet produce any too complex insights. However, the most important output would be to tangibly demonstrate the importance of the topic and lay some solid foundations for future work.
Photrek has proven its capability to deliver projects in this domain as well as in this funding format, and there are good reasons to believe that the persons with the assigned responsibilities are able to implement the plan also from a technical viewpoint.
Overall
5.0
5.0
4.0
GNU OSS license is great!
The team is uniquely qualified to implement a set of tooling for further study of alternative voting methods. However, as part of their expertise and prior insights they may also have pre-conceived notions about the range of possible algorithmic choices, and strong preferences.
I would prefer to relax the quadratic [square root] assumption and make it a tunable parameter rather than to hard code code it into the system. After many dozens of individual votes and over a year of conversations at Gitcoin I find the extra cognitive load distracting, while not necessarily producing desired outcomes. Meanwhile, a desire for uncorrelated votes may paint as nafarious collusion the outcomes of natural coordination and communication requisite in any functiouning organization.
A majority of token-based group decisions are of a binary approval, as can be seen on Snapshot. While aggregating the scores of 1-10 is a frequent enough method, and currently in use at Deep Funding, perhaps we should not be hard code that either.
While the averaged DF votes apparently center somewhere around 6 or 7, the frequency of the choices of 1 - 10 is somewhat similar, with incremental increases around the extremes. Yet, the instances of voting mostly 1 and only a single 10 is strongly frowned upon as 'strategic'. Not a question of a stochastic simulation, but can we look into normalizing the votes from a given reviewer toward whatever distribution is desirable, and perhaps decriminalize a choice of some voting distribution? However, current rules institute a minimum passing score, which is not immediately compatible with a simplistic renormalization.
The proof of the results of a simulation is backtesting the strategy suggested and/or making predictions and measuring the outcomes. Deep Funding does not yet have quantitative measures of past proposal and grant outcomes. When variously adjusting the ways to aggregate the votes, how do we compare different sets of parameters?
Incremental weighting adjustments, including taking a square root of some factor, will often enough leave the rankings of multiple choices largely intact. The data set of the votes and the outcomes of maybe a hundred proposals and under a hundred reviewers is unlikely to distambiguate fully the impact of a particular value of some parameter. A voting system proves its worth when the stakes are high, the decisions are complex, and the issues impactful and controversial. From my standpoint, so far the process has been polite and placid, so a study of its outcomes may not be fully informative.
The 'square root' thinking comes from the desire to depower the 'whales' and make decisions more democratic. However, the founders or early investors often have non-trivial token positions. Some Reputation factor aims to give additional voting power to the long-term participants and perhaps positive contributors. Already highly visible and engaged, they have outsize popularity and influence. Amplifying their input even further is a strong centralizing influence, which may be directly at odds with the democratizing aim of the 'square root' adjustment.
My apologies for having turned this review into a long exposition of personal opinions.
Funding put in reserve to address unforeseen requirements
Contract Signature
$2,400 USD
* Literature review for informed research design. * Prototyping Plural Voting: initial design a model to simulate multi-agent interactions aiming to analyze social phenomena and voting patterns in groups. * Monte Carlo computational modeling: simulate decisive social complex phenomena. * Selection of 2-3 experts by the community to form the ‘Red Team’
* A comprehensive literature review to guide the research design focusing on voting dynamics and individual preferences. * An Initial prototype of a Plural Voting model designed to uncover voting patterns and facilitate expressions in interactive groups. * A Monte Carlo computational model using random number sequences to model social voting dynamics. * Selected expert reviewers to form the ‘Red Team’.
$10,970 USD
* Quadratic Voting and No-Cost Grading implementated in computer simulations. * Comparison of the voting dynamics and outcomes between with and without quadratic costs for conviction voting. * Red team review of Photrek’s M2 deliverables.
* Fully implemented simulations of quadratic voting and No-cost Grading. * Complete and detailed design of the plural voting model capable of realistically simulating multi-agent interactions and social. * Comparison of Plural Voting with No-Cost Grading. * Comprehensive evaluation by red team experts of the plural voting prototype design and Monte Carlo modeling as presented in Milestone 2.
$11,080 USD
* Analysis of voting design modeling and simulation outputs utilizing careful statistical analysis including statistical moments (e.g. second moment kurtosis) * Provide summary of the relative performance outcomes of quadratic and no-cost grading. * Develop a proposal plan for how correlation measurements can be incorporated into the simulation for collusion mitigation. * Community Engagement: Red team review of Photrek’s M3 deliverables.
* Extensive data for knowledge extraction conclusions and decision-making support for the SNET community. * A comprehensive report featuring in-depth statistical analysis valuable insights and conclusions along with a detailed examination of complex system statistics. * A proposal plan for integrating correlation measurements into collusion mitigation simulations. * Comprehensive evaluation by red team experts of the quadratic voting implementation and complete plural voting design as presented in Milestone 3.
$11,660 USD
* Present a combination of visual analytical and statistical analyses. * Communicate key aspects of our design and research findings to the SNET community. * Offer insights and a potential roadmap for voting design improvements based on the simulation results. * Create a strategy that involves incorporating reputation scores into the simulation framework. Red team will provide independent recommendations based on the simulation results.
* Presentation of visual analytical and statistical research conclusions engaging the SingularityNET community with key findings and a roadmap for enhancements to SingularityNET voting platform. * Presentation of independent evaluations and recommendations from the red team.
$13,890 USD
Please create account or login to post comments.
Reviews & Ratings
New reviews and ratings are disabled for Awarded Projects
Overall
4.0
4.0
5.0
Desirability
The proposal addresses a critical need within SingularityNet's decentralized Deep Funding program by developing an open-source simulator to model the complex network of community sentiments and its impact on voting outcomes. The project's focus on evaluating different voting methods, such as plural voting and no-cost grading, to maximize the quality of community decisions is highly desirable. Additionally, the incorporation of Monte Carlo simulation techniques and statistical analysis adds rigor to the evaluation process. The proposal outlines clear objectives and milestones, indicating a strong commitment to advancing SingularityNet's governance capabilities. Further clarification on the specific methodologies for modeling community sentiments and implementing voting simulations would enhance the desirability of the project.
Viability
The project's viability for SingularityNet AI is evident in its potential to enhance the decision-making process within the Deep Funding program. By developing a simulation tool to evaluate different voting methods and their impact on community decisions, the project aligns with SingularityNet's mission of democratizing AI. The proposed milestones and deliverables align well with the project's objectives, demonstrating a practical approach to implementation. However, the proposal could provide more insights into the scalability of the simulation tool and its long-term impact on SingularityNet's AI ecosystem.
Feasibilty
The proposal demonstrates feasibility through a well-structured plan and budget allocation across five milestones. Milestones such as literature review, prototyping of Monte Carlo simulations, and analysis of voting dynamics provide a clear roadmap for project execution. The team's expertise in statistical modeling and simulation, as evidenced by their previous contracts with IOTA and Cardano Catalyst, adds credibility to their ability to carry out the project successfully. However, the feasibility could be further enhanced by providing more details on the technical specifications of the simulation tool and the integration process with SingularityNet's platform.
Usefulness
The proposal offers significant usefulness to SingularityNet AI by addressing critical issues related to decision-making and governance within the Deep Funding program. The project's focus on evaluating different voting methods, incorporating community sentiments, and providing insights into voting dynamics is invaluable for improving the quality and fairness of community decisions. The open-source nature of the simulator also allows for transparency and collaboration within the SingularityNet community. The proposal demonstrates a clear understanding of SingularityNet's objectives and presents a valuable initiative to support its growth and development.
Overall
5.0
5.0
5.0
The proposal targets a highly undervalued research area that is crucial for the survival of any intelligent self-organizing system including Deep Funding and SingularityNET. For this ecosystem, it is not a must-have project quite yet but it will be at some point in the near future and it is definitely not too early to start working on it.
These pretty straightforward simulations won’t yet produce any too complex insights. However, the most important output would be to tangibly demonstrate the importance of the topic and lay some solid foundations for future work.
Photrek has proven its capability to deliver projects in this domain as well as in this funding format, and there are good reasons to believe that the persons with the assigned responsibilities are able to implement the plan also from a technical viewpoint.
Overall
5.0
5.0
4.0
GNU OSS license is great!
The team is uniquely qualified to implement a set of tooling for further study of alternative voting methods. However, as part of their expertise and prior insights they may also have pre-conceived notions about the range of possible algorithmic choices, and strong preferences.
I would prefer to relax the quadratic [square root] assumption and make it a tunable parameter rather than to hard code code it into the system. After many dozens of individual votes and over a year of conversations at Gitcoin I find the extra cognitive load distracting, while not necessarily producing desired outcomes. Meanwhile, a desire for uncorrelated votes may paint as nafarious collusion the outcomes of natural coordination and communication requisite in any functiouning organization.
A majority of token-based group decisions are of a binary approval, as can be seen on Snapshot. While aggregating the scores of 1-10 is a frequent enough method, and currently in use at Deep Funding, perhaps we should not be hard code that either.
While the averaged DF votes apparently center somewhere around 6 or 7, the frequency of the choices of 1 - 10 is somewhat similar, with incremental increases around the extremes. Yet, the instances of voting mostly 1 and only a single 10 is strongly frowned upon as 'strategic'. Not a question of a stochastic simulation, but can we look into normalizing the votes from a given reviewer toward whatever distribution is desirable, and perhaps decriminalize a choice of some voting distribution? However, current rules institute a minimum passing score, which is not immediately compatible with a simplistic renormalization.
The proof of the results of a simulation is backtesting the strategy suggested and/or making predictions and measuring the outcomes. Deep Funding does not yet have quantitative measures of past proposal and grant outcomes. When variously adjusting the ways to aggregate the votes, how do we compare different sets of parameters?
Incremental weighting adjustments, including taking a square root of some factor, will often enough leave the rankings of multiple choices largely intact. The data set of the votes and the outcomes of maybe a hundred proposals and under a hundred reviewers is unlikely to distambiguate fully the impact of a particular value of some parameter. A voting system proves its worth when the stakes are high, the decisions are complex, and the issues impactful and controversial. From my standpoint, so far the process has been polite and placid, so a study of its outcomes may not be fully informative.
The 'square root' thinking comes from the desire to depower the 'whales' and make decisions more democratic. However, the founders or early investors often have non-trivial token positions. Some Reputation factor aims to give additional voting power to the long-term participants and perhaps positive contributors. Already highly visible and engaged, they have outsize popularity and influence. Amplifying their input even further is a strong centralizing influence, which may be directly at odds with the democratizing aim of the 'square root' adjustment.
My apologies for having turned this review into a long exposition of personal opinions.
© 2025 Deep Funding
HenriqC
Feb 10, 2024 | 3:42 PMEdit Comment
Processing...
Please wait a moment!
In my view, in the web 3 ecosystems, the real discussion about decision-making resources has barely even started. Yet the dynamic stability and evolution of these collective agents will be completely dependent on how the decision-making power is distributed among the subagents for each decision and over time. Objective metrics and fact based analysis are the only genuine guidance one can have to tackle such questions. Any error correction that can be done through this simulated ‘virtual reality’ means less cost on the actual process. I find this as a super important challenge and your proposal as an equally valuable endeavor.