Project Name: Pravahini (प्रवाहिनी) - Where Data Flows, Models Grow, and AI Glows!
Project Track: AI
Team Name: Team Pravahini
Team Member(s): 2 (@im_31 @victorious)
HackerEarth Project Link: HackerEarth Submission
Project Goal:
Our goal is to create a user-friendly and secure AI marketplace that facilitates the buying and selling of datasets and pre-trained ML models. Our platform also offers coding capabilities for the users and also empowers decentralized computation to process their machine learning models. We aim to establish a vibrant community of data providers, model developers, and users, offering a diverse range of high-quality resources for various domains using blockchain technology.
Through Pravahini, we aim to foster collaboration, innovation, and knowledge-sharing among data enthusiasts and researchers, driving ground-breaking advancements in AI and data-driven applications.
Project Value:
Benefits of Pravahini on Blockchain:
Secure Marketplace Functionality: Blockchain ensures a secure and transparent marketplace for buying and selling datasets and pre-trained ML models. Transactions are tamper-proof, enhancing trust.
Efficient Decentralized Computation: Leveraging blockchain, Pravahini optimizes ML model processing through the Bacalhau Protocol. Users harness the power of distributed systems for resource-intensive tasks.
Granular Access Control: The Lighthouse Protocol enforces secure access control, allowing users to define permissions. This protects sensitive data and resources from unauthorized access.
Transparent Revenue Sharing: Smart contracts facilitate transparent revenue sharing with data providers and model creators. This creates a fair and collaborative ecosystem, incentivizing quality contributions.
Project Info:
In Sanskrit, the term “प्रवाहिनी” (Pravāhinī) is used to describe a “flow” or “stream”. In the context of our platform it represents the continuous flow of data and models available for machine learning and artificial intelligence applications. It signifies the abundance and ongoing evolution of resources in the AI field. Pravahini is an innovative platform that offers an AI marketplace for datasets and pre-trained machine learning models and decentralized computation to process ML models.
Pravahini provides decentralized computation capabilities, allowing users to harness the power of distributed systems to process their machine learning tasks efficiently using our decentralized computation feature. In addition, our platform offers built-in coding functionality, empowering users to write and test their machine learning algorithms seamlessly.
Pravahini Presentation Video:
Pitch Deck:
Pravahini.pdf (654.4 KB)
Project Website: Pravahini
Smart Contract links:
Contract1 | Contract2 | Contract3
Updated Smart Contract links (Deployed on BitTorrent Mainnet):
Contract1 | Contract2 | Contract3
Project Test Instructions:
-
Connect your wallet:-
Connecting your wallet is the first step to using our Platform. -
Get Started:-
Click on the ‘Get Started’ button to be redirected to the sign-up page. -
User Dashboard:-
The user dashboard will provide you with access to all your datasets and machine learning models. Additionally, it will display the purchased models and datasets for your convenience. -
Create Dataset & ML Model:
To create a dataset or ML Model , upload a file from your local device in the required format. Fill in all the necessary details and click the submit button. -
Dataset Marketplace & ML Model Marketplace:
Here you can find the list of datasets and ML Models which you can buy and use. There will be free as well as paid Datasets and ML Models. -
Decentralized Computation:
Here you can use decentralized computing for your ML model computations. To facilitate this, we provide two robust containers tailored to meet your specific requirements.
Project Details:
Problem Statement
Existing marketplaces for AI resources lack a unified platform that seamlessly integrates dataset and pre-trained model trading, coding capabilities, and decentralized computation. This fragmentation hinders efficient collaboration and resource utilization within the AI community.
Solution
Unified Marketplace:
- Integrate dataset and pre-trained model trading on a single, intuitive platform.
- Streamline resource discovery, evaluation, and acquisition for users.
Decentralized Computation:
- Harness the power of distributed systems for efficient and scalable ML model processing.
- Enable users to leverage decentralized computation for resource-intensive tasks.
Judging criteria and how we fare
Originality: Pravahini redefines AI collaboration through a unique marketplace, integrating datasets, models, and decentralized computation with blockchain.
Technicality: Utilizing React.js, Bootstrap, Node.js, Express.js, Solidity, Lighthouse, and Bacalhau Protocols, Pravahini ensures a robust, scalable, and secure platform.
Business Model: Pravahini’s value lies in its unified marketplace, revenue sharing, competitive pricing, and premium features, ensuring a sustainable revenue stream.
Usability: With an intuitive UI and clear instructions, Pravahini abstracts blockchain complexities for seamless accessibility, even for non-experts.
Practicality: Addressing real-world data and computation needs, Pravahini empowers users across industries, leveling the playing field for all.
Contribution to the Crypto Ecosystem: Pravahini boosts crypto adoption, inviting new users, fostering collaboration, and driving innovation in AI and blockchain, offering substantial value to the community.
Team
im_31
- Role: Front-end Developer, AI Enthusiast
- Bio: A dedicated Front-end Developer with a passion for creating visually appealing and intuitive user interfaces. Actively interested in AI and committed to staying updated with the latest developments in the field to contribute to projects at the highest level.
Victorious
- Role: Front-end and Blockchain Developer
- Bio: A skilled Front-end Developer with a keen eye for visually appealing and user-friendly interfaces. Values creating websites that precisely cater to end-users’ needs. Combines expertise in web design with a passion for photography. Additionally, serves as a Blockchain Developer, contributing to the platform’s technological foundation.
Project Milestones:
Sprint 1 (25/07 - 11/08):
Frontend Tasks:
- Create the github repo and connect the Github repo with the Vercel App.
- Create the Navbar and Footer.(Responsive)
- Design and implement the Home Page with the project tagline and “Get Started” button.
- Implement Metamask integration to prompt users to connect their wallets when clicking “Get Started.”
- Create the Registration Page for users to provide Name, Occupation, Organization, Location, and upload Image for registration.
- Implement redirection logic based on wallet connection and registration status when clicking “Get Started” in the Home Page.
- Implement the User Dashboard with the Logo with and “Create” button in the Navbar/Sidebar.
- Implement the Datasets, Ml Models, Subscriptions Page.
- Create all the pages responsive and on each button add hover effect, cursor pointer and add toasters.
Backend (Smart Contract) Tasks:
- Implement the smart contract function for user authentication and authorization.
- Implement the smart contract functions to allow users to create datasets with Title, Description, Categories, Price per data, License, Dataset file, and Image.
- Implement smart contract functions to set dataset visibility (Public, Private, or Sell) and record the settings upon dataset creation.
- Implement smart contract functions to search and retrieve dataset details based on specific criteria.
- Implement smart contract functions to handle dataset purchases and transfer payments to the dataset owner.
- Testing of the functions created for datasets.
Sprint 2 (12/08 - 30/08):
Frontend Tasks:
- Implement the Dataset Marketplace section with Upload Dataset, Search Bar, and display of Free/Paid datasets.
- When someone uploads the dataset, set the Plagiarism Checker with the Datasets available on the platform.
- Implement logic to view individual dataset and in the single dataset view the required data.
- In the single dataset apply logic for the button to be able to download the datasets in the different formats - json, csv, xlsx - after making the payment.
- Implement the Model Marketplace section with Upload Model, Search Bar, and display of Free/Paid models.
- Implement logic to view individual Model and model details when clicking “View More” buttons.
- Implement the transaction process when clicking the “Buy” button for models and datasets.
- In Models Marketplace after making the payment the user must be able to download the ML Model.
- Create all the pages responsive and on each button add hover effect, cursor pointer and add toasters.
Backend (Smart Contract) Tasks:
- Implement the smart contract functions to allow users to create ML models with Title, Description, Categories, Tags/Keywords, License, Model file, and Usage documentation.
- Implement smart contract functions to set model visibility (Public, Private, or Sell) and record the settings on model creation.
- Implement smart contract functions to search and retrieve model details based on specific criteria.
- Implement smart contract functions to handle model purchases and transfer payments to the model owner.
- Testing of the functions created for ML Models.
Sprint 3 (31/08 - 18/09):
Frontend Tasks:
- Implement the “Code” and front-end for “Decentralized Computation” sections to provide code execution and decentralized ML model computation.
- Design and implement the display of user addresses and their Content IDs (CIDs) in a table.
- Create all the pages responsive and on each button add hover effect, cursor pointer and add toasters.
- Complete all the remaining and pending work for frontend.
Backend (Smart Contract) Tasks:
- Complete the remaining Work.
- Integration of Smart Contract with Front-end
- Implement the smart contract functions to handle subset creation for datasets (if required).
- Optimize the smart contract for gas efficiency and overall performance.
- Conduct comprehensive testing of the frontend and smart contract functionalities to ensure proper functionality and security.
Sprint 4 (18/09 - 26/09):
- Complete the Content of the Platform.
- Create the presentation of the Platform.
- Complete all the remaining work for front-end as well as back-end.
Project Screenshots:
