A comment on Guo et al. [arXiv:2206.11228]
This is an incremental commentary that addresses methodological concerns in neuroscience and machine learning research, potentially affecting researchers in those fields.
The authors critique a prior study by Guo et al., which claimed that adversarially trained neural networks achieve robustness comparable to primate IT neural representations, by questioning the interpretation and phrasing of these results without providing new experimental data or alternative findings.
In a recent article, Guo et al. [arXiv:2206.11228] report that adversarially trained neural representations in deep networks may already be as robust as corresponding primate IT neural representations. While we find the paper's primary experiment illuminating, we have doubts about the interpretation and phrasing of the results presented in the paper.