papercopilot/paperlists
Processed / Cleaned Data for Paper Copilot
What it solves
Paper Lists provides a centralized, consistent repository of AI conference paper records. It solves the problem of fragmented information where paper data is scattered across OpenReview, official conference sites, and open access platforms, making it difficult for researchers to navigate and analyze conference trends.
How it works
The project aggregates data from multiple sources into structured JSON files for major AI conferences (such as ICLR, NeurIPS, ICML, CVPR, EMNLP, and CoRL). It includes a local search tool built with Streamlit that allows users to search and analyze these records locally via a web interface or a command-line interface (CLI) to extract specific fields based on keywords.
Who it’s for
AI researchers and data analysts who need a comprehensive, structured dataset of conference papers for statistical analysis or efficient local searching of academic literature.
Highlights
- Comprehensive Coverage: Includes historical data for major conferences like NeurIPS (dating back to 2000) and ICLR.
- Structured Data: Provides records in JSON format for easy programmatic access.
- Local Search Tool: A Streamlit-based app and CLI tool for keyword-based extraction and analysis.
- Multi-source Integration: Combines data from official sources and open access sites to ensure coherence.