DataTerminology.comNEW
A platform linking machine learning papers with implementations, datasets, tasks, and benchmarks.
DESCRIPTION
Papers with Code organizes machine learning research by connecting papers to code repositories, datasets, evaluation tasks, and leaderboards. It helps researchers compare state-of-the-art results, reproduce experiments, and find open-source implementations. The platform curates tasks across domains, supports standardized metrics, and tracks progress over time. Integration with arXiv and GitHub enables timely updates and collaboration.
ALTERNATIVE DEFINITIONS
Papers with Code
We create a free and open resource with machine learning papers, code, datasets and evaluation tables.
Emphasizes openness and the linkage of papers to code and datasets.
Meta AI
Papers with Code is a free and open resource for researchers and practitioners to find machine learning papers, code, datasets, and benchmarks.
Highlights its role for both researchers and practitioners and benchmark tracking.
RELATED TERMS
Benchmark Leaderboards Dataset Reproduction State of the Art Code Repository
COMPETING TERMS
OpenReview arXiv
KNOWLEDGE AREAS
AI Data Science Platform & Tooling Collaboration
TARGET AUDIENCE
Data Scientist AI Engineer MLOps Engineer AI Product Manager Data Analyst
EXTERNAL SOURCES
- Wikipedia - Papers with Code
- Papers with Code - Primary platform and documentation
- OpenReview - Peer-review platform for ML papers
- arXiv - Preprint repository for research
TOP QUESTIONS
What is Papers with Code used for?
Papers with Code helps find ML papers linked to open-source implementations, datasets, tasks, and leaderboards. It supports reproducibility and comparison of state-of-the-art results.
Category: Overview
How does Papers with Code track the state of the art?
It maintains task-specific leaderboards with standardized metrics. Authors and the community submit results verified against benchmark datasets.
Category: Management
Why is Papers with Code relevant for Data Scientists?
It accelerates experimentation by providing code references and benchmark baselines, reducing time to reproduce and extend methods.
Category: Relevancy
How does Papers with Code integrate with arXiv?
Papers with Code ingests arXiv identifiers and metadata to link papers to code, datasets, and tasks. This ensures timely updates and discoverability.
Category: Integration
What are common mistakes when using Papers with Code?
Relying solely on leaderboard scores without reading methods or data details can mislead. Always confirm evaluation protocols and dataset splits.
Category: Mistakes
What risks exist when following Papers with Code leaderboards?
Overfitting to benchmarks and ignoring real-world constraints may occur. Ensure robust validation and consider domain-specific performance.
Category: Risk
What are alternatives to Papers with Code?
OpenReview and arXiv provide paper discovery, while model hubs like Hugging Face host implementations and datasets. Each emphasizes different aspects.
Category: Alternatives
How do I contribute to Papers with Code?
Submit results, code links, or datasets via the platform’s contribution forms. Ensure clear documentation, reproducible scripts, and valid metrics.
Category: How-To
Where should beginners start with Papers with Code?
Explore task pages for curated benchmarks, read linked papers, and run provided implementations on sample datasets to understand workflows.
Category: Getting-Started