# Comparing representations of biological data learned with different AI paradigms, augmenting and cropping strategies

arXiv: https://arxiv.org/abs/2203.04107 · HTML: https://mlsanity.com/paper/2203.04107/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 12 of its 30 references that are in the map.
- First posted on arXiv 7 Mar 2022.

## Builds on (most cited first, 12 of 12)

- [A Simple Framework for Contrastive Learning of Visual Representations](https://mlsanity.com/paper/2002.05709/) (2020)
- [Bootstrap your own latent: A new approach to self-supervised Learning](https://mlsanity.com/paper/2006.07733/) (2020)
- [UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction](https://mlsanity.com/paper/1802.03426/) (2018)
- [Unsupervised Learning of Visual Features by Contrasting Cluster Assignments](https://mlsanity.com/paper/2006.09882/) (2020)
- [Improved Training of Wasserstein GANs](https://mlsanity.com/paper/1704.00028/) (2017)
- [Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks](https://mlsanity.com/paper/1511.06434/) (2015)
- [Automatic chemical design using a data-driven continuous representation of molecules](https://mlsanity.com/paper/1610.02415/) (2016)
- [Data Efficient and Weakly Supervised Computational Pathology on Whole Slide Images](https://mlsanity.com/paper/2004.09666/) (2020)
- [Unsupervised Learning for Cell-level Visual Representation in Histopathology Images with Generative Adversarial Networks](https://mlsanity.com/paper/1711.11317/) (2017)
- [Evaluation Metrics for Unsupervised Learning Algorithms](https://mlsanity.com/paper/1905.05667/) (2019)
- [Tune It or Don't Use It: Benchmarking Data-Efficient Image Classification](https://mlsanity.com/paper/2108.13122/) (2021)
- [On denoising autoencoders trained to minimise binary cross-entropy](https://mlsanity.com/paper/1708.08487/) (2017)

## Abstract (arXiv)

Recent advances in computer vision and robotics enabled automated large-scale biological image analysis. Various machine learning approaches have been successfully applied to phenotypic profiling. However, it remains unclear how they compare in terms of biological feature extraction. In this study, we propose a simple CNN architecture and implement 4 different representation learning approaches. We train 16 deep learning setups on the 770k cancer cell images dataset under identical conditions, using different augmenting and cropping strategies. We compare the learned representations by evaluating multiple metrics for each of three downstream tasks: i) distance-based similarity analysis of known drugs, ii) classification of drugs versus controls, iii) clustering within cell lines. We also compare training times and memory usage. Among all tested setups, multi-crops and random augmentations generally improved performance across tasks, as expected. Strikingly, self-supervised (implicit contrastive learning) models showed competitive performance being up to 11 times faster to train. Self-supervised regularized learning required the most of memory and computation to deliver arguably the most informative features. We observe that no single combination of augmenting and cropping strategies consistently results in top performance across tasks and recommend prospective research directions.