Validators selection

• Team Name: Optymalizacja AI Grzegorz Miebs
• Payment Address: bc1qtcjq0jpcup43ny5e66f6kuvcn9pyhamguecsgu BTC
• Level: 1

Project Overview 📄

Response to an RFP validator-selection-algorithm.

Overview

This project aims to create a decision-support tool aiding nominators in selecting validators based on their individual preferences. These preferences are expressed in a very easy and intuitive way by performing pairwise comparisons. A nominator has to answer several times (around 6) which one out of two present validators they prefer. Based on these comparisons a mathematical model reflecting the nominator's preference is created. Finally, the model is used to rank validators. I was already involved in a research phase of this project hence I'd like to make a final version.

Project Details

The aim of this project is only a backend. The final result will be a Python flask application exposing its functionality via RESTful API

• Functionality:
  • Providing a pair of validators for comparison:
    • Input:
      • previous comparisons
    • Output:
      • next pair
      • current model’s quality
      • current model
  • Providing a ranking for a given model
    • Input:
      • model
    • Output:
      • ranking of validators
  • Accepting new data
    • Input:
      • validators.csv file that contains information of recent era data from trusted sources

Ecosystem Fit

This application will be used in a validators selection phase, thus all nominators are its audience. The project makes the selection process easier and more robust. To the best of my knowledge, there isn't a similar project.

Team 👥

Team members

• Grzegorz Miebs

Contact

• Contact Name: Grzegorz Miebs
• Contact Email: grzegorz.miebs@protonmail.ch
• Website:

Legal Structure

• Registered Address: Poland, Poznan 61-853, Wierzbowa 2/22
• Registered Legal Entity: Optymalizacja AI Grzegorz Miebs

Team's experience

I have 4 years of industry experience as a data scientist and 6 years of academic experience in a multicriteria decision support field. The most relevant project is of course study regarding this topic with the preprint available here: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4253515 Other related projects:

• Multicriteria job planning for bus and tram drivers for the public transport system of Poznań
• Selection of a sustainable third-party reverse logistics provider https://doi.org/10.1016/j.omega.2018.05.007

Team Code Repos

• https://github.com/miepsik

Team LinkedIn Profiles

• https://www.linkedin.com/in/grzegorz-miebs/

Development Status 📖

• preprint https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4253515

Development Roadmap 🔩

Overview

• Total Estimated Duration: 30 days
• Full-Time Equivalent (FTE): 1FTE
• Total Costs: 9,000 USD

Milestone 1 Example — Basic functionality

• Estimated duration: 4 weeks
• FTE: 20 days
• Costs: 6,000 USD

Number	Deliverable	Specification
0a.	License	Apache 2.0
0b.	Documentation	We will provide both inline documentation of the code and a basic tutorial that explains how a user can (for example) spin up one of our Substrate nodes and send test transactions, which will show how the new functionality works.
0c.	Testing and Testing Guide	Core functions will be fully covered by comprehensive unit tests to ensure functionality and robustness. In the guide, we will describe how to run these tests.
0d.	Docker	We will provide a Dockerfile(s) that can be used to test all the functionality delivered with this milestone.
0e.	Article	We will publish an article/workshop that explains how this algorithm works and how to use the software
1.	Next pair	Develop an algorithm for efficient calculations of the next pair to be compared to maximize the model’s information gain.
2.	Ranking calculation	Develop an algorithm calculating a score for each validator
3.	New data	Develop a function for the data preprocessing
4.	Internal testing	Unit tests covering the functionality and logic

Milestone 2 (Testing)

• Estimated Duration: 2 weeks
• FTE: 10 days
• Costs: 3000 USD

Number	Deliverable	Specification
0a.	License	Apache 2.0
0b.	Documentation	We will provide both inline documentation of the code and a basic tutorial that explains how a user can (for example) spin up one of our Substrate nodes and send test transactions, which will show how the new functionality works.
0c.	Testing and Testing Guide	Core functions will be fully covered by comprehensive unit tests to ensure functionality and robustness. In the guide, we will describe how to run these tests.
0d.	Docker	We will provide a Dockerfile(s) that can be used to test all the functionality delivered with this milestone.
0e.	Article	We will publish an article/workshop that explains how this algorithm works and how to use the software
1.	Deployment	Deploy the code on a test server provided by the Grants Team or by myself.
2.	Test live environment	Test the server efficiency by checking an average response time for each endpoint and provide a report
3.	Polishing	Reach out for feedback to the Grants Team. Integrate final feedback on functional, as well as cosmetic changes like the way data are provided, configuration, etc.

Future Plans

The possible extensions are:

• analysis of the obtained results and nominators' preferences
• capturing drift of preferences and just updating the model instead of repeating the whole pairwise comparison procedure

Additional Information ➕

How did you hear about the Grants Program? Personal recommendation