Computer Vision News - August 2021

2 Summary AI Rese ch 4 are thankful to all the authors ( Arunachalam Narayanaswamy, Subhashini Venugopalan, Dale R. Webster, Lily Peng, Greg Corrado, Paisan Ruamviboonsuk, Pinal Bavishi, Rory Sayres, Abigail Huang, Siva Balasubramanian, Michael Brenner, Philip Nelson, Avinash V. Varadarajan ) for allowing us to use their images to illustrate this review and especially to Subha for providing us with extra illustrations and links. Their paper can be found here. Every month, Computer Vision News selects a research paper to review. This month’s article is dedicated to Scientific Discovery by Generating Counterfactuals Using Image Translation , written by several authors from the Google Health research centre and in collaborationwith the Rajavithi Hospital. This work was presented at MICCAI 2020 and it is based on preliminary work which appeared in Nature Communications at the start of last year! We Scientific Discovery by Generating Counterfactuals Using Image Translation by Marica Muffoletto The scope of this paper, just like for many other papers these days, is to increase our knowledge of what causes the predictions of a deep learning model. As visual recognition models are more and more employed, especially in medical imaging (where the tasks are more complex and require specific knowledge), the establishment of excellent explanation techniques is becoming crucial. This work aims to go beyond explanation techniques to provide means for scientific discovery. How to do that? By finding out what about the data leads to the predictive power of a classification model, rather than where in the image the model focuses to build its predictions.