: NLP Vocabulary: probability distribution

NLP methods and tools > statistics > probability distribution

probability distribution

Adaptive-Consistency models the probability distribution over unique samples using a Dirichlet distribution allowing us to quantify the confidence in the lead of the majority element over other elements. (Aggarwal, Madaan, Yang & Mausam, 2023)
As such this problem can be solved naturally with optimal transport (OT) methods that facilitate the computation of the optimal mapping between two probability distributions. (Phung, Minh Tran, Nguyen & Nguyen, 2021)
BERT is a large Transformerbased language model that has achieved strong performance in many NLP tasks and DistilBERT uses a process known as knowledge distillation to reproduce its behavior by training a smaller model to replicate its probability distributions across class predictions. (Farinango Cuervo & Parde, 2022)
Thus the statistical distribution of the words observed in training data has a crucial role to guide the NMT models. (Ataman & Federico, 2018)
We then use these two probability distributions to calculate the Vaserstein distance (Vaserstein 1969) thus obtaining the edge displacement Vaserstein distance (EDV) for a given dataset. (Anderson & Gómez-Rodríguez, 2022)

http://data.loterre.fr/ark:/67375/8LP-BNJWDW78-1

RDF/XML TURTLE JSON-LD Last modified 6/13/24

Loterre