Deep Funding Round 5: Proposed Changes and Community Feedback Process

As we move toward Deep Funding Round 5 (DFR5), we’re presenting a set of suggested changes designed to improve the process for proposers, reviewers, and the SingularityNET (SNET) ecosystem as a whole. These proposed changes are based on feedback from 114 community members, including expert reviewers and proposers from previous rounds. The suggestions have been discussed extensively in 10 weekly workgroup meetings involving key stakeholders.

However, these changes are not final. We are opening them up for community debate and feedback. Based on the input we receive during this second round of feedback, Deep Funding and All-Circles will assess and adapt the proposals to create the final version of changes for Round 5.

In this blog post, we’ll outline the proposed changes and explain the next steps for how you can provide feedback through the link at the bottom of this article.

Workgroup and Changes Creation Process

The suggestions for Round 5 are based on comprehensive community engagement, including feedback collected through three surveys in English and one in Spanish. The surveys asked community members about what to keep, what to change, and what to stop.

The workgroup tasked with proposing changes included:

  • Deep Funding Staff
  • Marketing Circle Member
  • Review Circle Member
  • Expert Reviewer
  • DFR4 Proposer
  • UX/UI Development Team

The workgroup dedicated over 100+ hours in total, working across 10 weekly meetings and numerous asynchronous discussions on Mattermost. Each member contributed an average of 20+ hours to ensure the feedback was properly assessed and actionable changes were proposed.

The top issues identified included the quality of proposals, reviews, and voting—and these areas are the primary focus for improvement in DFR5.

Proposed Changes for Round 5

Here’s a breakdown of the key changes proposed for DFR5:

  1. Pools for Proposals

The proposed change in pools for Deep Funding Round 5 aimed at streamlining resource allocation and improving the quality of funded projects, while sticking to the program’s purpose which is to onboard quality services on the SNET’s Marketplace. This pool change is designed to ensure that Deep Funding focuses its resources on high-quality, competitive services while providing better guidance and support for proposers.

Specifically, the change focuses on reducing the type of proposals submitted categorizing them across only two pools:

  • AI Services Pool:
    • Proposers can submit projects that will:
      • Be developed from scratch.
      • Be adapted for onboarding to the SingularityNET marketplace.
      • Be connected to onboarded projects to consume existing services.
      • Improve an existing onboarded service.
  • Ideation Pool:
    • It will address key aspects such as:
      • Marketing Research: Understanding the market demand and positioning for the AI service.
      • Financial Planning: Ensuring the project has a sound financial strategy for execution and sustainability.
      • Technical Overview: Detailing the technical requirements and approach for developing the AI service.
      • Onboarding Process: Planning how the service will be integrated into the SingularityNET marketplace.
  • Former Pools:
    • Pools such as marketing and tooling will move under a Request for Proposals (RFP) framework. For example, the marketing pool will start with an RFP for Regional Town Halls, which is an example of how more targeted proposals will be submitted for specific tasks.
    • Should community members have different ideas from the ones proposed by Df or SNET, they will still be able to propose RFPs under a new process flow that will be available later.
  1. Proposal Submission 

Proposal and Reviewer Categorization

Proposals and reviewers will be categorized by AI field with the proposals also falling under industry categories. This will allow experts to review proposals in their field of expertise while allowing voters to also assess the proposals by their industry of choice.

We are also introducing Marketing Experts into this mix, allowing the scoring and selection process of awarded teams to be reliant on another important dimension for the success of services in the market.

The key AI fields include:

  • Natural Language Processing (NLP)
  • Computer Vision
  • Forecasting
  • Reinforcement Learning
  • Data Science and Analytics
  • Neuro-symbolic AI

For industry categories, we are sticking with the previous list as the following:

  • Agriculture & Food
  • Algorithmic/technical
  • Commerce
  • Community and Collaboration
  • Finance/DeFi
  • Health and wellness
  • HR
  • Language
  • Learning and education
  • Media & Entertainment
  • Productivity
  • Robotics
  • Safety and ethics
  • Social welfare
  • Sustainability & Environment
  • Transportation & Mobility
  • Inclusion
  • Other

Proposal Milestones and Success Criteria

In DFR5, expert reviewers will help proposers set clear and specific milestones with success criteria indicators. This allows for better tracking of project progress and alignment with Deep Funding goals, ultimately leading to more successful outcomes.

  1. Expert Review and Review Process

Expert Reviewers Involved Throughout the Proposal Lifecycle

One of the key changes proposed for Deep Funding Round 5 is the extended role of expert reviewers, who will now be involved throughout the entire lifecycle of a proposal. This means that from the initial submission through to the final delivery and onboarding of the service on SingularityNET’s marketplace, the expert reviewers will play an active role in guiding and reviewing the projects – including their milestone deliverables.

Gated Review Process

A more structured review process will be introduced to improve the quality of proposals:

  • Eligibility Review: Initial proposal submissions will be audited by the designated expert reviewer to ensure they meet basic criteria.
  • Expert Review:
    • 80% of the scoring will come from human expert reviewers.
    • 20% of the scoring will be handled by AI (ChatGPT), using standardized prompts to ensure fairness and consistency.
    • The prompt will be created by expert reviewers individually, concatenated by the DF staff, with the main sentences, and kept confidential until the final scores.
    • Proposals that score 7 or above will proceed to the next phase – Interview.
  • Interview Phase: Proposals that advance will undergo live interviews with a team of experts to ensure that proposers have the necessary expertise to execute the project.
  • Appeal Phase: Teams will be able to appeal the outcome of their interviews, providing an opportunity for reconsideration.
  • Community Voting: The new Photrek algorithm introduced in Round 4, will be used after a second round of tests and iteration.

Improved Expert Reviewer Selection

A dedicated Expert Reviewer Lead will assess current reviewers to determine if they qualify for expert status. New reviewers will be selected based on their expertise and academic peer-reviewing experience.

This system will ensure that only highly qualified individuals are conducting the reviews, leading to better outcomes for proposers and the program.

Systematic Proposal Distribution

To prevent bias, proposals will be assigned to reviewers in a specific order, without the option of choice. This ensures that no reviewer can selectively choose only the highest-potential projects, resulting in a fairer review process.

  1. Veto Power for Deep Funding

The introduction of veto power will allow Deep Funding to veto any proposal that does not align with its strategic objectives, even after the final reviews. This measure ensures that funded projects are in line with the long-term vision of the program.

  1. Communication and Inquiry Handling

To improve communication between proposers and Deep Funding, a dedicated Telegram channel will be set up. This will serve as a centralized platform for inquiries, ensuring that questions are answered quickly and consistently. This channel aims to foster better engagement and transparency throughout the proposal process.

  1. Proposal Limits

Proposers and teams will be limited to a maximum of 3 proposals or awarded projects in total for any given funding round. This change is designed to ensure a balanced workload for proposers, improve the overall quality of submissions, and facilitate more efficient project execution within the Deep Funding program.

Why This Change?

  • Focus on Project Delivery: Proposers with ongoing awarded projects will be encouraged to focus on delivering their projects instead of submitting new proposals, improving the chances of timely and successful project completion.
  • Encourages Higher Quality Proposals: With a cap on the number of submissions, proposers will need to prioritize their best ideas, resulting in higher-quality proposals.
  • More Focus per Proposal: By reducing the number of allowed submissions, proposers will have more time to develop and follow through on each proposal, improving the overall quality and success rate of the projects.
  • Faster Project Execution: Once a proposer completes their ongoing projects, they can submit new ones, speeding up execution and allowing them to propose again sooner.
  1. Timeline for Round 5
  • Weeks 1-4: Submissions open. Proposers can choose to make their proposals visible or hidden during this time. Eligibility review, expert review, and feedback begin simultaneously.
  • Week 5: Eligibility review continues, and feedback is provided to teams. Adjustments to proposals can be made based on the feedback.
  • Week 6: Final eligibility list is published, and expert reviews take place.
  • Week 7: Proposals are reviewed and scored (80% by expert reviewers, 20% by ChatGPT, with a standardized prompt for all). Top-scoring proposals (7+) proceed to the interview phase. Interviews for AI Services over $50K are conducted, and feedback reports are provided.
  • Week 8: Appeal period for declined interviews; final list of approved proposals is prepared.
  • Week 9: Awareness campaign for approved proposals begins. Transition to the voting period.
  • Week 10: Voting period begins.
  • Weeks 11-12: Voting closes; results are calculated and announced.

Next Steps: Community Feedback Process

The proposed changes for Deep Funding Round 5 are currently open for community feedback. Here’s the process moving forward:

  • Until Oct 3rd: Community members can submit feedback through the feedback form (link below).
  • Assessment: All feedback will be reviewed by Deep Funding and All-Circles.
  • Adaptation: Based on the input received, the proposed changes will be adjusted to reflect the community’s feedback.
  • Presentation: The final version of the changes will be shared through the appropriate channels.
  • Timeframe: The final version is expected to be ready within one to two months, depending on the community’s desire for deeper discussions on the changes.

 

Provide Your Feedback

Your input is crucial to the success of DFR5. To share your thoughts on the proposed changes, please fill out the feedback form:

Community Feedback Form