Skip to content

Latest commit

 

History

History
105 lines (70 loc) · 4.54 KB

README.md

File metadata and controls

105 lines (70 loc) · 4.54 KB

Paperbaum - Parachain for Academic Papers

Paperbaum is a decentralized academic paper publishing and verification system built on a custom Substrate-based parachain. It addresses issues of authorship verification, restricted access, and inefficient paper linking in academic publishing.

Link to the slides: here

Core Components

Screenshot 2024-07-21 at 11 23 39

  • Substrate Parachain: Custom runtime for paper metadata storage in a merkle tree and verification.
  • IPFS Integration: Decentralized storage for full paper content.
  • Vector Similarity Engine: NLP-based system for semantic paper linking.

Technical Features

Substrate Parachain

The core of Paperbaum is built on a custom Substrate parachain, providing a robust and flexible foundation for academic paper management and verification. The custom pallet uses a Merkle tree to natively link papers together. This pallet provides functionality for:

  • Managing a Merkle tree of paper hashes
  • Verifying Merkle proofs
  • Storing and retrieving paper metadata
  • Enforcing size limits on various paper attributes

IPFS Integration

Paperbaum leverages the InterPlanetary File System (IPFS) for decentralized storage of full paper content. This integration ensures that papers are stored in a distributed, content-addressed manner, enhancing accessibility and permanence.

Vector Similarity Engine

Paperbaum implements a vector similarity engine for semantic paper linking. This system uses OpenAI's text embedding model to generate vector representations of papers, enabling efficient similarity searches. The generateEmbedding function creates a vector representation of text, while cosineSimilarity computes the similarity between two vectors.

Paper Processing and Metadata Extraction process

When a paper is uploaded, Paperbaum processes the PDF, extracts key metadata, and generates a vector representation:

  1. PDF text extraction
  2. Metadata extraction using GPT4o-mini
  3. Vector embedding generation
  4. IPFS upload
  5. Storage of metadata and vector in-memory in a merkle tree

Images

How to Use

Parachain

For the parachain, first compile it using

cargo build --release

and then run

./target/release/node-template --dev

to run the substrate node on 127.0.0.1:9944

Backend

To run the backend server, enter the backend directory and run

npm install

and then proceed to run

node server.js

to run the server on localhost:3000

Frontend

To run the frontend, enter the frontend directory and run

npm install

and then

npm run dev

to run the frontend on localhost:3001

Roadmap

1. Enhanced Merkle Tree Implementation

  • Develop a more sophisticated Merkle tree structure for efficient paper linking and verification.
  • Implement Merkle Mountain Ranges (MMR) for dynamic dataset management, allowing efficient updates and proofs of inclusion.

2. Implement zk-SNARKs for privacy-preserving paper submissions and verifications.

  • Develop a ZK-based reputation system for anonymous yet credible peer reviews.
  • Create ZK proofs for citation verification without revealing full paper contents.

3. Design a token-based incentive mechanism for peer reviewers.

  • Implement double-blind review processes using ZK proofs.
  • Develop a reputation system for reviewers based on the quality and timeliness of their reviews.

4. Implement XCM for communication with other parachains.

  • Develop cross-chain citation verification and tracking.
  • Create a system for recognizing academic credentials and reputations across different blockchain networks.

5. Use OriginTrail's Decentralized Knowledge Graph (DKG) for semantic linking of academic papers.

  • Implement versioning and provenance tracking of papers using OriginTrail's blockchain-agnostic protocol.
  • Develop an AI-assisted discovery system leveraging OriginTrail's semantic data structure.

LICENSE

See MIT License