Dear Editor,<\/p>\n
In a reply to our letter \u201cCOVID-19, AI enthusiasts, and toy datasets: radiology without radiologists\u201d [1], S\u00e9rgio D. Dias and Pedro E. M. Lopes (FastCompChem, Lda) \u201cadvocate a two-way communication between the clinical and technological worlds.\u201d We certainly support this suggestion. The critique in our letter is directed toward works that exactly do not value this \u201ccommunication\u201d and as non-radiologists publish sketchy results based on toy datasets in violation of all established good practices for curating reliable datasets. Dias and Lopes state that their \u201cview aligns in general with the opinion\u201d expressed in our letter. However, they also say that in their \u201cexperience, toy datasets are still of value, even when not following all accepted standards.\u201d This could be a major concern when works get published using toy datasets amid a pandemic with aggrandizing names such as \u201cCOVID-Net\u201d with implied or direct claims on a diagnosis. It seems Dias and Lopes would agree with this as well. They continue that they \u201cused the compilation of Cohen et al. \u2026 as source of radiographs. The dataset is certainly not the well curated\u2026but it was quickly available allowing the initial development effort.\u201d Although there may be some value in using toy datasets for initial developments in the lab, what we categorically reject is reporting \u201cfaulty results based on creating amateur datasets and training sketchy AI solutions hastily to publish online.\u201d Later improvements of such faulty results and publishing them in reputable places do not help either because the damage has already been done.<\/p>\n
Dias and Lopes \u201cfully agree with the ideas\u201d we expressed in our letter, but they \u201cwould not discard the use of toy datasets if that is the only avenue of development, including training of deep learning algorithms.\u201d Initial experiments in a lab may be fine with toy datasets. However, training deep networks with amateur data is risky if those networks are made publicly available. We take comfort in Dias and Lopes further agreement when they state: \u201cthe true validation can only occur in controlled clinical trials under the supervision of radiologists though.\u201d<\/p>\n
<\/p>\n","protected":false},"excerpt":{"rendered":"
Dear Editor, In a reply to our letter \u201cCOVID-19, AI enthusiasts, and toy datasets: radiology without radiologists\u201d [1], S\u00e9rgio D. Dias and Pedro E. M. Lopes (FastCompChem, Lda) \u201cadvocate a two-way communication between the clinical and technological worlds.\u201d We certainly support this suggestion. The critique in our letter is directed toward works that exactly do […]<\/p>\n","protected":false},"author":3,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_et_pb_use_builder":"","_et_pb_old_content":"","_et_gb_content_width":"","inline_featured_image":false,"footnotes":""},"categories":[6],"tags":[],"class_list":["post-1039","post","type-post","status-publish","format-standard","hentry","category-opinions"],"acf":[],"yoast_head":"\n