Français
5. Generative models
A generative model is an algorithm that generates samples according to the same probability law as the data provided in the initial database, but without this probability law necessarily being explicitly described.
5.1 Variable autoencoders
We have seen how deep learning methods can be used to perform dimensionality reduction using auto-encoders by associating with each object x i in a very high-dimensional space a latent representation h i = f θ (x), where f θ is a neural network, and by optimizing the parameters θ of f θ and...
The Ultimate Scientific and Technical Reference
This article is included in
Technological innovations
This offer includes:
Knowledge Base
Updated and enriched with articles validated by our scientific committees
Services
A set of exclusive tools to complement the resources
Practical Path
Operational and didactic, to guarantee the acquisition of transversal skills
Doc & Quiz
Interactive articles with quizzes, for constructive reading
Generative models
Conclusion
Bibliography
Software tools
For computations that do not involve deep learning and that deal with data volumes that do not require the use of distributed computing, the two reference software tools are scikit-learn and R
The Spark Mlib library adapts the main machine learning algorithms (excluding deep learning) to a distributed environment, enabling the processing of very large volumes of data.
Deep...
Events
Annual conferences :
International Conference on Learning Representations ( https://iclr.cc/ )
Conference on Neural Information Processing Systems ( https://nips.cc/ )
Conference on Computer Vision and...
The Ultimate Scientific and Technical Reference
