# In Defense of the Paper

arXiv: https://arxiv.org/abs/2104.08359 · HTML: https://mlsanity.com/paper/2104.08359/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 20 of its 51 references that are in the map.
- First posted on arXiv 16 Apr 2021.

## Builds on (most cited first, 20 of 20)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [A Style-Based Generator Architecture for Generative Adversarial Networks](https://mlsanity.com/paper/1812.04948/) (2018)
- [You Only Look Once: Unified, Real-Time Object Detection](https://mlsanity.com/paper/1506.02640/) (2015)
- [Continuous control with deep reinforcement learning](https://mlsanity.com/paper/1509.02971/) (2015)
- [Playing Atari with Deep Reinforcement Learning](https://mlsanity.com/paper/1312.5602/) (2013)
- [Sequence to Sequence Learning with Neural Networks](https://mlsanity.com/paper/1409.3215/) (2014)
- [Extracting Training Data from Large Language Models](https://mlsanity.com/paper/2012.07805/) (2020)
- [Energy and Policy Considerations for Deep Learning in NLP](https://mlsanity.com/paper/1906.02243/) (2019)
- [Soft Actor-Critic Algorithms and Applications](https://mlsanity.com/paper/1812.05905/) (2018)
- [Deep Reinforcement Learning that Matters](https://mlsanity.com/paper/1709.06560/) (2017)
- [Distributional Reinforcement Learning with Quantile Regression](https://mlsanity.com/paper/1710.10044/) (2017)
- [Improving Reproducibility in Machine Learning Research (A Report from the NeurIPS 2019 Reproducibility Program)](https://mlsanity.com/paper/2003.12206/) (2020)
- [Differential Privacy Has Disparate Impact on Model Accuracy](https://mlsanity.com/paper/1905.12101/) (2019)
- [Characterising Bias in Compressed Models](https://mlsanity.com/paper/2010.03058/) (2020)
- [Maxmin Q-learning: Controlling the Estimation Bias of Q-learning](https://mlsanity.com/paper/2002.06487/) (2020)
- [Troubling Trends in Machine Learning Scholarship](https://mlsanity.com/paper/1807.03341/) (2018)
- [Fairwashing: the risk of rationalization](https://mlsanity.com/paper/1901.09749/) (2019)
- [The Radicalization Risks of GPT-3 and Advanced Neural Language Models](https://mlsanity.com/paper/2009.06807/) (2020)
- [The Mirage of Action-Dependent Baselines in Reinforcement Learning](https://mlsanity.com/paper/1802.10031/) (2018)
- [HARK Side of Deep Learning -- From Grad Student Descent to Automated Machine Learning](https://mlsanity.com/paper/1904.07633/) (2019)

## Abstract (arXiv)

The machine learning publication process is broken, of that there can be no doubt. Many of these flaws are attributed to the current workflow: LaTeX to PDF to reviewers to camera ready PDF. This has understandably resulted in the desire for new forms of publications; ones that can increase inclusively, accessibility and pedagogical strength. However, this venture fails to address the origins of these inadequacies in the contemporary paper workflow. The paper, being the basic unit of academic research, is merely how problems in the publication and research ecosystem manifest; but is not itself responsible for them. Not only will simply replacing or augmenting papers with different formats not fix existing problems; when used as a band-aid without systemic changes, will likely exacerbate the existing inequities. In this work, we argue that the root cause of hindrances in the accessibility of machine learning research lies not in the paper workflow but within the misaligned incentives behind the publishing and research processes. We discuss these problems and argue that the paper is the optimal workflow. We also highlight some potential solutions for the incentivization problems.