Глоссариум по искусственному интеллекту: 2500 терминов. Том 2

The adjusted predictions and probabilities should match the distribution of an observed set of labels[190 - Calibration layer [Электронный ресурс] https://developers.google.com URL: https://developers.google.com/machine-learning/glossary#iteration (дата обращения: 02.05.2023)].

Candidate generation — the initial set of recommendations chosen by a recommendation system[191 - Candidate generation [Электронный ресурс] https://developers.google.com URL: https://developers.google.com/machine-learning/recommendation/overview/candidate-generation (дата обращения: 10.

01.2022)].

Candidate sampling is a training-time optimization in which a probability is calculated for all the positive labels, using, for example, softmax, but only for a random sample of negative labels. For example, if we have an example labeled beagle and dog candidate sampling computes the predicted probabilities and corresponding loss terms for the beagle and dog class outputs in addition to a random subset of the remaining classes (cat, lollipop, fence).

The idea is that the negative classes can learn from less frequent negative reinforcement as long as positive classes always get proper positive reinforcement, and this is indeed observed empirically. The motivation for candidate sampling is a computational efficiency win from not computing predictions for all negatives[192 - Candidate sampling [Электронный ресурс] https://developers.google.com URL: https://developers.google.com/machine-learning/glossary#candidate-sampling (дата обращения: 28. 03.2023)].

Canonical Formats in information technology, canonicalization is the process of making something conform] with some specification… and is in an approved format. Canonicalization may sometimes mean generating canonical data from noncanonical data. Canonical formats are widely supported and considered to be optimal for long-term preservation[193 - Canonical Formats [Электронный ресурс] www.umich.edu URL: https://www.icpsr.umich.edu/web/ICPSR/cms/2042#C (дата обращения: 07.

07.2022)].

Capsule neural network (CapsNet) is a machine learning system that is a type of artificial neural network (ANN) that can be used to better model hierarchical relationships. The approach is an attempt to more closely mimic biological neural organization[194 - Capsule neural network [Электронный ресурс] https://ru.what-this.com URL: https://ru.what-this.com/7202531/1/kapsulnaya-neyronnaya-set.html (дата обращения: 07.02.