This is the very first Expert review, created by a team of 4 team members on projects that were awarded in Round 1
This review is part of an experiment where we are asking existing teams to help our community assess the quality of submitted proposals.
As part of this experiment, we collected 8 volunteering team members and divided them into 2 groups of 4.
Each group chose a Pool that they wanted to review To avoid potential conflicts of interest, we added the restriction that a reviewer will not assess a Pool in which they have submitted a proposal themselves. This has led to the selection of Pool B: existing services and Pool C; Ideation projects being assessed as part of this first experiment. The goal is, as the pool of teams will grow over the following rounds to have all pools and proposals assessed in this way.
About the process:
As this is an experiment there is no fixed process, and the teams have some freedom to organize themselves. There are however a few guidelines:
Group process:
We assign value to the group process. Discussing one’s assessment in a group with peers offers an opportunity to explore and discover new insights and align differentiated views. We also believe that this process is an extra (future) safeguard against favoritism. Of course, we have full confidence in the teams that are currently doing the assessments but this is also part of the process-design for future rounds with potentially larger groups of assessors and less visibility on the participants.
Assessment criteria.
We asked the team members to evaluate each project against 3 criteria:
- Feasibility (is it technically possible to make this?)
- Viability (will this team, with this budget be able to accomplish this with sufficient quality in a reasonable time?)
- Desirability (Will the platform benefit? Will it help the platform grow? Is it aligned with the values of SNET? Does it offer good value for money?)
The teams are making these assessments first on an individual basis and compare their findings in a group process.
Below, find the recommended projects in pool C – Projects in their ideation phase and, further below, in Pool B – Existing Services.
Assessment group 1 – Projects in Pool C – Projects in their ideation phase
Including the grades and detailed assessment results by the assessment team:
Project: Research Guild
Ratings:
| Feasibility |
4 |
| Viability |
5 |
| Desirability |
5 |
| Average |
4.7 |
Interesting project with broad application, reasonable funding request, specific prototype deliverable, specific proposal to create a marketplace service as a ResearchAI portal using customized ChatGPT portal that specializes in generating academic level research reports through an appealing UI (subsequent funding round.)
The proposed project of using AI tools to support academic research by increasing collation, analysis and reporting efficiencies and reducing the time taken to do so seems useful. The integration of AI solutions with existing research processes can potentially lead to improved research outcomes for industries by providing more up-to-date and comprehensive insights in a more efficient manner. AI has the potential to automate repetitive tasks and provide advanced analysis capabilities, freeing up researchers to focus on more creative and strategic aspects of their work.
To make the proposed project viable and feasible, the following recommended steps:
- Market research: Conduct a thorough market analysis to determine the demand for such a solution and to assess the potential competition in the field.
- Resource allocation: Ensure that sufficient resources, including budget and human capital, are allocated for the project to maximize its chances of success.
- Project planning: Develop a clear project plan that outlines the timeline, milestones, and resources needed for each stage of the project.
- Technical feasibility: Conduct a technical feasibility study to assess the feasibility of integrating AI solutions with existing research processes and to anticipate any potential limitations or challenges that may arise.
- Team building: Build a strong and experienced team, including developers with AI/UI skillsets, to ensure the success of the project.
- Prototyping: Develop a clickable UI prototype to demonstrate the potential of the solution and to secure additional funding or support.
- Continuous evaluation: Regularly evaluate the progress of the project and make necessary adjustments to ensure its success.
By taking these steps, the proposed project can increase its chances of becoming both viable and feasible.
Project: VeterinaryDAO
Ratings:
| Feasibility |
5 |
| Viability |
4 |
| Desirability |
5 |
| Average |
4.7 |
The VeterinaryDAO is seeking funding to help produce a proposal in the next round for an AI Assisted Veterinary Triage system. The team has well articulated needs for creating a successful proposal in the next round and clearly stated deliverables in this ideation phase, including market research to identify and testing existing online veterinary triage services, developmental work into options for developing an AI service for veterinary triage assistance as well as SingularityNET AI marketplace services that could benefit veterinary triage assistance. These milestone deliverables are tangible and demonstrable. This is well within the scope of an ideation phase project in that it will (publish or) utilize a service on the SNET platform.
The team is capable, with experience in veterinary medicine and marketing. The budget is modest. An implementation of this system will likely require additional staffing.
Project: The creator economy gamified and tokenized
Ratings:
| Feasibility |
4 |
| Viability |
3 |
| Desirability |
4 |
| Average |
3.7 |
SoundDAO is proposing to gamify the experience of creating NFT content. The game is an IRL P2E environment. It would be helpful to add the full phrase In-real-life Play-to-Earn. The proposal describes players progressing from a Street team to a CEO, though I’m not sure “becoming a CEO” is an attractive goal for an artist. Perhaps the journey could be described in a more inclusive manner of different kinds of accomplishments.
Nine milestones ranging from defining the minimum viable product (MVP) to launch and evaluation of the MVP are described. There is no breakdown of the $10,000 budget, which makes it difficult to evaluate the viability of the project achieving its goals. The proposal specifies that more personnel beyond the current team, CEO and COO, are needed but the additional roles are not described.
Assessment group 2 – Pool B – Existing Services
Purpose of our meeting was to discuss our evaluations of projects we felt warrant a ‘yes’ recommendation. Primary criteria are how the proposals grow the marketplace, and how does it progress SNET towards a beneficial AI (AGI).
Project: Simulating Risky worlds
Easily the most technically-prepared proposal is Simulating Risky Worlds, put forth by Photrek. We’ve made notes that another evaluation by a machine learning professional would be welcome. However, we’re also highly confident in the Photrek team because of their skill, knowledge, and diligence demonstrated in their DeepFunding Round 1 proposal.
We’ve reached out to the Photrek team representative for clarification on the service, and whether it is meant to work on purely mathematical models, virtual simulation, real-world topographical data, or all of the above and more.
Photrek’s reply:
Q: Will the tech work in 3d simulations?
A: In principle, our approach could work in 3D. We’ll probably start with 2D to evaluate the performance.
Q: The service will test ML weights to identify weaknesses?
A: Photrek’s innovation is measuring the Robustness of forecasts and using that to optimize the learning not just for accuracy but Robustness against Outlier cases.
Q: The simulations capability will be a service?
A: Yes, it will be a service. We’ll have to train offline, so the service will allow generation of simulation samples from an example we train in advance.
Project: Upgrade of the Stable Diffusion (SD) Service
Our second project evaluation is for Upgrade of the Stable Diffusion (SD) Service. This was an easy ‘yes’ for our recommendation. The author clearly has experience with the current SNET SD API, and clearly identified areas where friction was encountered in their own experience, and which functions appeared to be missing. It’s clear that generative art is very popular now; an improved API could greatly benefit SNET’s marketplace development.
Project: Onboard NeuralProphet: a hybrid time-series forecasting library
Our third proposal, Onboard Neuroprofit, was also an easy recommendation. The authors propose a time-series prediction service for SNET, something many SNET projects would utilize.
Post Views: 747