ML-Powered NFT Recommendation API, Trained on the NuNet Network

[GENERAL] Name and surname of main applicant

Anthony Capitan

[GENERAL] Email address of main applicant

brerrabbit3@protonmail.com

Additional applicants

• Giovanni Gargiulo 
• Chadi Nassar 
• Ilija Radeljic
• Kabir Veitas

[GENERAL] Please specify how many months you expect your project to last (from 2-12 months)

10

[GENERAL] Please indicate if your proposal has been auto-translated into English from another language.

No

[GENERAL] Does your project have any dependencies on other organizations, technical or otherwise?

No

[GENERAL] If YES, please describe what the dependency is and why you believe it is essential for your project’s delivery. If NO, please write “No dependencies.” .

No dependencies.

[GENERAL] Will your project’s output/s be fully open source?

No

[GENERAL] If NO, please describe which outputs are not going to be open source. If YES, please write “Project will be fully open source.”

ArgusNFT will be open-sourcing code pertaining to the training pipeline of our ML model - starting from the pre-processing of the dataset using the NuNet network, through to the full, model training pipeline. This code will no doubt be beneficial to subsequent projects who leverage it to train their own ML/AI models using NuNet.

And as a open-source project, NuNet will be publishing all project outputs. 

That said, there are two elements of the NFT Recommender API that will not be open-source:

1. Dataset creation/extractio
2. Data hydration: while running inference (running the model) will happen on the nunet network following industry standard practices, results (that will be policy id + asset names) will have to be hydrated to all the NFT data, and this hydration process will not be open-souce

[METADATA] Category of Proposal

NFTs

[METADATA] SDG rating

SDG 9: Industry, Innovation and Infrastructure

[IMPACT] Please describe your proposed solution.

Our proposed solution is an NFT recommendation API that can be integrated by any Cardano NFT platform. The API will use machine learning (ML) to analyze the transaction history of a user's wallet and generate a personalized selection of NFTs that are listed for sale (or rent) on the platform. The API will leverage NuNet's decentralized computing network to pre-process our data set and train the ML model.

We believe that this solution will address the problem of lackluster NFT search and discovery, which leads to missed sales opportunities and low user retention for NFT marketplaces. By providing users with relevant and appealing NFT suggestions, we aim to increase their engagement, satisfaction, and loyalty. We also hope to attract new users who are looking for a more personalized and curated NFT experience.

Our solution is unique because it combines the power of (public) wallet data, NuNet’s decentralized computing network, and our extensive experience implementing ML solutions in ecommerce – to bring an ML strategy to the Cardano NFT market that has already proven revolutionary for Web2 applications. We will also integrate our API with Plutus.Art, a leading Cardano NFT marketplace, as a proof of concept and a showcase for this solution’s potential. And we’ll work with a SaaS attorney to draft a licensing framework that will allow other platforms to easily implement the API at minimal cost.

[IMPACT] How does your proposed solution address the challenge and what benefits will this bring to the Cardano ecosystem?

ML-based Recommender Systems are defining features of virtually every leading, Web2 commerce and content application. From the video recommendations users see on YouTube, to the product recommendations they’re presented with on Amazon, ML-based recommendations play and integral role in the UX of the world’s leading platforms. But to our team’s knowledge, recommender systems are virtually non-existent in the NFT space.

This is especially strange given that the personalized marketplace experience made possible recommender systems is such a great fit for NFT marketplaces: Public blockchain provide ample data for the development of effective recommender systems. And NFT marketplaces typically sell a diverse array of assets whose value/appeal is highly subjective, and thus perfectly suited to personalized UX based on individual user preferences.

That said, our Recommendation API will benefit the Cardano ecosystem in a number of ways:

1) Increasing NFT Marketplace Transactions:

A 2021 Barilliance study (https://www.barilliance.com/personalized-product-recommendations-stats/) found that ecommerce platforms which incorporated ML-based, product recommendation systems could attribute an average of 12% of their sales to recommendations, and that some platforms were getting as much as 31% of their sales from recommendations. Of course, product recommendations can come in some forms that are outside the scope of this current proposal (such as on-platform and email notifications). But even a more modest gain of increasing marketplace sales by 5% (which is our mid-term goal for the NFT recommendation API) is very significant if the solution is widely adopted.

2) Facilitating cooperative models and increased sustainability between NFT platforms: 

For example, NFT analytics sites will be able to link their users directly to marketplaces where they can buy recommended NFTs, thus making both analytics sites and marketplaces more sustainable. 

3) It will showcase the potential of NuNet as a decentralized computing network that can support ML-based applications on Cardano.

This is a new use case for the NuNet network, and this project will accelerate the growth of NuNet’s capacity to support additional AI/ML projects. 

4) Moving towards more personalized UX/UI on NFT platforms will benefit a broader diversity of NFT projects and users. 

Currently, UI on leading NFT platforms typically centers around trading volume “leaderboard stats” and thumbnails from the most popular NFT collections on a blockchain. 

The absence of a personalized, NFT discovery experience on NFT platforms is the Web3 equivalent of being advertised Mr. Beast videos every time you sign onto YouTube, even though you only ever watch videos on how to paint watercolors. And this suboptimal UX not only affects customer retention and sales. It also undermines the success of a broader diversity of creators, by needlessly allotting all the UI “real estate” to projects with the most popular appeal in the current market. 

[IMPACT] How do you intend to measure the success of your project?

We intend to measure the success of our project using both quantitative and qualitative metrics. Some of the metrics we will use are:

• The number of platforms that integrate our API
• The number of users who connect their wallets to our API
• The number of NFT recommendations generated by our API
• The conversion rate of recommendations (how many users buy or rent the recommended NFTs)
• The retention rate of users who use our API (how often they come back to use it again)
• The satisfaction rate of users who use our API (how much they like or dislike the recommendations)
• The feedback from platforms and users on our API (what they like or dislike about it, what they suggest for improvement)

We will collect these metrics using various tools such as analytics software, surveys, interviews, and social media. We will also compare these metrics with baseline data from before we launch our API to measure the impact and improvement. We will report these metrics to the Catalyst community and the Cardano Foundation on a regular basis.

We estimate that our solution will impact at least 10% of the current Cardano NFT users within one year of launching our API, based on our market research and user feedback. We also expect to attract new users who are interested in personalized NFT recommendations. We aim to reach at least 5 Cardano NFT platforms that will integrate our API within one year of launching it.

[IMPACT] Please describe your plans to share the outputs and results of your project?

ArgusNFT and Plutus.Art activity:

We plan to share the outputs and results of our project in various ways, such as:

• Publish recommendation-api integration guides
• Open source the model training pipeline on the NuNet network
• Create examples/guides/video
• Integrating our API with Plutus.Art, a leading Cardano NFT marketplace, and inviting users to try it out and provide feedback.
• Create a blog entry and website to showcase our project, its features, its benefits, and its impact.
• Promoting our project on social media platforms such as Twitter, Reddit, Telegram, and Discord, and engaging with the Cardano NFT community.
• Participating in relevant events, webinars, podcasts, and interviews to present our project and its results.
• Reaching out to other Cardano NFT platforms and offering them to license and integrate our API at minimal.
• Reporting our progress, achievements, challenges, and learnings to the Catalyst community and the Cardano Foundation on a regular basis.

Partner NuNet activity

1. Spreading Outputs Over a Timescale

Our project plan includes clear milestones and deliverables, which will be shared publicly as they are completed. This incremental release of outputs will ensure a continuous stream of updates for the community.

This approach lets us provide updates on a regular basis, and offers users the chance to provide feedback that we can use to guide subsequent development. We 

2. Sharing Outputs, Impacts, and Opportunities

We intend to leverage NuNets and ArgusNFTs various communication channels to share our project's outputs, impacts, and opportunities:

• GitLab: The primary hub for our technical work, hosting our codebase, documentation, and issue tracking. This will be the main point of reference for the details of our project.
• Social Platforms: We plan to regularly post updates on our progress on platforms like Twitter, Discord, LinkedIn, and Reddit. This will include major milestones, bug fixes, and insights from our development work.
• Technical Discussions: We will continue to hold weekly technical discussion where we discuss the technical aspects of our work. This provides a forum for live Q&A and discussion with our community. 
• Blogs: A regular blogs to summarize the progress we have made, highlighting key achievements and outlining the next steps in our project.

3. Testing and further research

As an open-source project, our outputs will be freely accessible for further research and development. We encourage the community's involvement in testing our solutions to enhance their real-world performance.

Community Testing: We'll invite our users to participate in alpha and beta testing phases, where they can help identify bugs and suggest improvements. We'll use GitLab's issue tracking for managing feedback and provide guidelines for issue reporting and feature suggestions.

Internally, we'll use project insights and community feedback to guide our future work, optimize performance, and prioritize new features. Our aim is to foster a collaborative development ecosystem that is robust, relevant, and of high quality.

[CAPABILITY/ FEASIBILITY] What is your capability to deliver your project with high levels of trust and accountability?

Our team has the capability to deliver this project with high levels of trust and accountability, because:

• We have a strong background and experience in ML, NFTs, blockchain, web development, and SaaS.
• We have successfully completed previous projects related to Cardano NFTs, such as ArgusNFT (a Cardano NFT market analytics platform) and Plutus.Art (a Cardano NFT marketplace).
• We have successfully completed Project Catalyst proposals in Fund8 and Fund9, both on time and on schedule (ArgusNFT)
• We have established partnerships and collaborations with other projects in the Cardano ecosystem, such as NuNet (a decentralized computing network) and Project Catalyst (a community innovation platform).
• We have a clear vision, mission, and roadmap for our project, as well as a detailed budget breakdown and a risk management plan.
• We have a transparent and accountable approach to managing funds properly, such as using escrow services, smart contracts, audits, reports, and receipts.

[CAPABILITY/ FEASIBILITY] What are the main goals for the project and how will you validate if your approach is feasible?

• To develop an NFT recommendation API that uses ML, wallet data, and NuNet to personalize NFT search and drive sales.
• To integrate our API with Plutus.Art as a proof of concept and a showcase for other platforms.
• To license our API to other Cardano NFT platforms at minimal cost.
• To increase the sales, user retention, and competitive advantage of Cardano NFT platforms that use our API.
• To enhance the diversity, quality, and discoverability of Cardano NFTs.
• To showcase the potential of NuNet as a decentralized computing network that can support ML-based applications on Cardano.
• To promote the adoption of Cardano as a platform for NFT creation, exchange, and consumption.

[CAPABILITY/ FEASIBILITY] Please provide a detailed breakdown of your project’s milestones and each of the main tasks or activities to reach the milestone plus the expected timeline for the delivery.

• Milestone 1: Data collection and pre-processing using NuNet.
• Tasks:
• Collect wallet data from Cardano blockchain and NFT platforms using APIs and web scraping tools.
• Pre-process the data using NuNet's decentralized computing network to clean, transform, and normalize the data - Both of NuNet’s ML on CPU and ML on GPU platforms can be useful here.
•   Store the data in a secure and accessible database at Argus.
• Timeline: 1 month.

• Milestone 2: ML model development and training using NuNet.
• Tasks:
•   Design and implement an ML model that can generate personalized NFT recommendations based on wallet data.
• Containerize implemented ML applications per NuNet deployment requirements; 
•   Train and test the ML model using NuNet's decentralized computing network to optimize its performance and accuracy.
•   Evaluate and validate the ML model using various metrics and techniques.
• Timeline: 3 months.

• Milestone 3: API development and testing.
• Tasks:
•   Develop an API that can expose the ML model functionality to external platforms and users.
•   Test the API functionality, performance, security, usability, compatibility, and scalability using various tools and methods.
•   Document the API specifications, features, usage, and examples using clear and comprehensive documentation.
• Timeline: 2 months.

• Milestone 4: Integration with Plutus.Art.
• Tasks:
•   Integrate our API with Plutus.Art, a leading Cardano NFT marketplace, as a proof of concept and a showcase for other platforms.
•   Invite Plutus.Art users to connect their wallets to our API and receive personalized NFT recommendations based on their transaction history.
•   Collect feedback from Plutus.Art users on their experience and satisfaction with our API and its recommendations.
• Timeline: 2 months.

• Milestone 5: Licensing framework and documentation.
• Tasks:
•   Work with a SaaS attorney to draft the licensing framework and agreement for our API that will allow other platforms to license and integrate our API at minimal cost.
•   Document the licensing framework and agreement and provide clear and comprehensive information on how other platforms can license and integrate our API.
• Timeline: 1 month.

• Milestone 6: Marketing and outreach to other platforms.
• Tasks:
•   Promote our project and our API to the Cardano NFT community and other platforms using various channels and methods, such as website hosting, social media ads, events, webinars, podcasts, interviews, etc.
•   Reach out to other Cardano NFT platforms and offer them to license and integrate our API at minimal cost.
•   Provide support and guidance to other platforms that want to license and integrate our API.
• Timeline: 1 month.

[CAPABILITY/ FEASIBILITY] Please describe the deliverables, outputs and intended outcomes of each milestone.

Milestone 1:   Success criteria:

•   A large and diverse data set of wallet transactions and NFT listings is obtained and ready for ML model development.
•   The data quality and integrity are ensured and verified using NuNet's computing resources.
•   The data is stored in a secure and accessible database that can be queried and updated easily.

Deliverables:

Outputs: 

Milestone 2:  Success criteria:

•   A robust and reliable ML model that can generate personalized NFT recommendations based on wallet data is developed and trained.
•   The ML model achieves high performance and accuracy on various metrics and techniques, such as precision, recall, F1-score, A/B testing, etc.
•   The ML model is evaluated and validated using NuNet's computing resources and feedback from potential users.

Deliverables:

Outputs:

Milestone 3:  Success criteria:

•   An API that can provide personalized NFT recommendations based on wallet data to external platforms and users is developed and tested.
•   The API meets the requirements and expectations of functionality, performance, security, usability, compatibility, and scalability.
•   The API documentation is clear and comprehensive and covers all the aspects of the API specifications, features, usage, and examples.

Deliverables:

Outputs:

Milestone 4:  Success criteria:

•   Our API is successfully integrated with Plutus.Art and provides personalized NFT recommendations to its users based on their wallet data.
•   Plutus.Art users are able to connect their wallets to our API easily and receive relevant and appealing NFT suggestions.
•   Feedback from Plutus.Art users is positive and constructive and helps us improve our API and its recommendations.

Deliverables:

Outputs:

Milestone 5:  Success criteria:

•   A licensing framework and agreement for our API is drafted and reviewed by a SaaS attorney and meets the legal and ethical standards of SaaS products and services.
•   The licensing framework and agreement is documented and provides clear and comprehensive information on how other platforms can license and integrate our API.

Deliverables:

Outputs:

Milestone 6: Success criteria:

•   Our project and our API are widely known and recognized by the Cardano NFT community and other platforms.
•   At least 5 Cardano NFT platforms express interest in licensing and integrating our API within one year of launching it.
•   We provide support and guidance to other platforms that license and integrate our API and ensure a smooth and successful integration process.

Deliverables:

Outputs:

[RESOURCES & VALUE FOR MONEY] Please provide a detailed budget breakdown of the proposed work and resources.

• ML developer: $30,000. A full-time ML developer who will be responsible for data collection, pre-processing, model development, and training using NuNet.
• Web developer: $27,500. A full-time web developer who will be responsible for API development, testing, documentation, and integration with Plutus.Art.
• SaaS attorney: $10,000. A professional SaaS attorney who will help us draft the licensing framework and agreement for our API.
• NuNet credits: $20,000. The amount of credits we will need to use NuNet's decentralized computing network for our data and ML tasks.
• Marketing and outreach: $10,000. The expenses related to promoting our project and our API to the Cardano NFT community and other platforms, such as website hosting, social media ads, events, webinars, podcasts, interviews, etc.
• Project management: $5,000. The costs of managing the project, such as communication tools, software licenses, hardware, etc.
• Documentation: $3,000. The costs of creating clear and comprehensive documentation for our API, such as video tutorials, user guides, FAQs, etc.
• Reporting: $2,000. The costs of reporting our progress, achievements, challenges, and learnings to the Catalyst community and the Cardano Foundation, such as surveys, analytics software, etc.

Total Budget - $107,500

[RESOURCES & VALUE FOR MONEY] How does the cost of the project represent value for money for the Cardano ecosystem?

We believe that the cost of the project represents value for money for the Cardano ecosystem because:

• We have based our budget on realistic estimates of the time and resources required to complete this project with high quality standards.
• We have benchmarked our costs against the average wage in our countries of residence and the typical freelance rates in our industry (ML/web development/blockchain).
• We have considered alternative sources of funding or revenue for this project, such as grants, sponsorships, or fees, but we have concluded that Catalyst funding is the most suitable and sustainable option for us at this stage.
• We have demonstrated that our project will bring significant benefits and value to the Cardano ecosystem, such as increasing NFT sales, user retention, and adoption, showcasing NuNet's potential, enhancing NFT diversity, quality, and discoverability, and fostering collaboration and interoperability among NFT platforms.

[IMPORTANT NOTE] The Applicant agreed to Fund10 rules and also that data in the Submission Form and other data provided by the project team during the course of the project will be publicly available.

I Accept