Intelligent Recommendation and Indexing System for Deep Learning Papers

Demo Video

Refer to YouTube

Local Deployment

System Requirements

1. MySQL or MariaDB w./ schema dl2, on localhost:3306: systemctl enable --now mysqld
2. Neo4j on localhost:7474: <NEO4J_HOME>/bin/neo4j restart

Basic Services

1. User & paper service on localhost:8080: cd UserService;./gradlew run
2. Neo4j graph service on localhost:8081: cd Neo4jService;./gradlew run
3. Frontend service on localhost:4200: cd frontend;ng serve

Intelligent Services

1. Vectorized search service w./ weaviate on Docker: sudo docker run -d -p 8090:8090 -p 50051:50051 dl2-vectorized;cd SearchLambda;python serve_test.py
2. Gemini-based GraphRAG: cd DataAnalysis/llm;python graph_rag.py
3. GAT-based paper recommendation: cd DataAnalysis;python gat_pred.py

Crawler Services

1. Initialization w./ renowned PapersWithCode methods and tasks: cd CrawlerService;npm run init
2. Expand the graph w./ Neo4j uncharted papers, exported as CSVs: npm run import
3. Daily routine crawling w./ subscription to ArXiV: npm run subscribe