Login / Signup

Hidden Variables in Deep Learning Digital Pathology and Their Potential to Cause Batch Effects: Prediction Model Study.

Max SchmittRoman Christoph MaronAchim HeklerAlbrecht StenzingerAxel HauschildMichael WeichenthalMarkus TiemannDieter KrahlHeinz KutznerJochen Sven UtikalSebastian HaferkampJakob Nikolas KatherFrederick KlauschenEva Krieghoff-HenningStefan FröhlingChristof von KalleTitus Josef Brinker
Published in: Journal of medical Internet research (2021)
Because all of the analyzed hidden variables are learnable, they have the potential to create batch effects in dermatopathology data sets, which negatively affect AI-based classification systems. Practitioners should be aware of these and similar pitfalls when developing and evaluating such systems and address these and potentially other batch effect variables in their data sets through sufficient data set stratification.
Keyphrases
  • deep learning
  • electronic health record
  • big data
  • artificial intelligence
  • machine learning
  • primary care
  • human health
  • risk assessment
  • general practice