此參考說明所註明的目錄。請向該機構確認當前課程供給與適用於你入學的條件。
課程說明
機率圖模型與神經生成模型 介紹生成模型的理論基礎及其在預測、知識發掘與創意設計中的應用。機率圖模型（PGMs）的基礎、傳統 PGM（如隱馬可夫模型、混合模型與潛在狄利克雷分配）的學習與推論。無向神經生成模型（NGMs），包括馬可夫隨機場與波茲曼機變體。有向 NGMs，包括赫姆霍茲機與深度信念網、變分自編碼器與生成對抗網路。
原文參考文本
Probabilistic Graphical Models and Neural Generative Models Introduction to the theoretical foundations of generative models and their applications in prediction, knowledge discovery, and creative design. Foundations of probabilistic graphical models (PGMs), learning, and inference in traditional PGMs such as hidden Markov models, mixture models, and latent Dirichlet allocation. Undirected neural generative models (NGMs), including Markov random fields, and variants of Boltzmann machines. Directed NGMs, including Helmholtz machines and deep belief nets, variational autoencoders, and generative adversarial networks.
來源與參考
保留日期與來源以協助你核實資料。為便於閱讀提供譯文；官方來源為條件與要求的參照。
來源參考 : https://brocku.ca/webcal/2024/graduate/cosc.html