Papers
A curated list of the top 100 most cited deep learning papers from 2012-2016, serving as a foundational reading list.
What is Papers?
Awesome Deep Learning Papers is a curated GitHub repository listing the top 100 most cited academic papers in deep learning from 2012 to 2016. It serves as a foundational reading list to help researchers and engineers understand the seminal works that shaped the modern deep learning landscape. The project filters the vast literature to highlight papers with the highest impact and broadest applicability.
Deep learning researchers, graduate students, and practitioners who want a structured, high-quality entry point into the core literature of the field without being overwhelmed by its volume.
It provides a community-vetted, citation-based filter to identify the most influential papers, saving time and offering a clear learning path. Unlike broader paper lists, it focuses on a critical historical period and maintains a limited, manageable size to ensure quality over quantity.
Overview
The most cited deep learning papers
Use Cases
Best For
- Graduate students starting deep learning research and needing a reading roadmap
- Practitioners seeking to understand the foundational concepts behind modern models
- Researchers conducting literature reviews in core deep learning subfields
- Educators compiling reading lists for advanced machine learning courses
- Self-learners looking for a curated, historical perspective on deep learning breakthroughs
- Teams wanting to build shared knowledge on seminal works in computer vision or NLP
Not Ideal For
- Researchers or practitioners seeking the latest deep learning advancements published after 2016
- Teams needing hands-on code implementations or tutorials to accompany academic papers
- Individuals focused on highly specialized or emerging subfields not represented in the top 100 seminal works
- Projects requiring interactive or dynamically updated paper recommendations beyond a static list
Pros & Cons
Pros
The list selects papers based on explicit citation thresholds (e.g., +800 citations for 2012 papers), ensuring only high-impact works are included, as detailed in the 'Awesome list criteria' section of the README.
Papers are categorized into topics like Convolutional Networks and NLP/RNNs, making it easy to navigate specific areas of interest, as shown in the structured Contents section.
Includes sections for datasets, software, books, and tutorials, providing a holistic learning path beyond just paper lists, as seen in the HW/SW/Dataset and Book/Survey/Review sections.
Was actively maintained with community contributions and a process for adding/removing papers to keep the list at 100, though now archived, as mentioned in the contributing guide and update notes.
Cons
The project is no longer maintained since 2017 due to the overwhelming number of new papers, making it irrelevant for recent research developments, as stated in the notice at the top of the README.
Focuses only on papers from 2012-2016, so it misses both older foundational works (only briefly covered in 'Old Papers') and all post-2016 breakthroughs, which limits its current comprehensiveness.
It's purely a list of papers with links to PDFs; there are no accompanying code repositories or practical examples, which might hinder hands-on learning despite the fetch_papers.py script for downloading.
Frequently Asked Questions
Related Projects
Open Source Society University🎓 Path to a free self-taught education in Computer Science!
Awesome machine learningA curated list of awesome Machine Learning frameworks, libraries and software.
University Courses:books: List of awesome university courses for learning Computer Science!
Data Science:memo: An awesome Data Science repository to learn and apply for real world problems.
Found a gem we're missing?
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.