Key Features

• Implement a variety of deep-learning algorithms in Keras and notice how deep-learning can be utilized in games
• See how quite a few deep-learning types and sensible use-cases will be applied utilizing Keras
• A useful, hands-on consultant with real-world examples to provide you a robust origin in Keras

Book Description

This booklet starts off via introducing you to supervised studying algorithms reminiscent of uncomplicated linear regression, the classical multilayer perceptron and extra subtle deep convolutional networks. additionally, you will discover photo processing with reputation of hand written digit pictures, category of pictures into assorted different types, and complex gadgets attractiveness with comparable photo annotations. An instance of id of salient issues for face detection can be supplied. subsequent you can be brought to Recurrent Networks, that are optimized for processing series facts corresponding to textual content, audio or time sequence. Following that, you are going to know about unsupervised studying algorithms akin to Autoencoders and the highly regarded Generative opposed Networks (GAN). additionally, you will discover non-traditional makes use of of neural networks as kind Transfer.

Finally, you are going to examine Reinforcement studying and its software to AI video game taking part in, one other renowned path of analysis and alertness of neural networks.

What you are going to learn

• Optimize step by step features on a wide neural community utilizing the Backpropagation Algorithm
• Fine-tune a neural community to enhance the standard of results
• Use deep studying for photo and audio processing
• Use Recursive Neural Tensor Networks (RNTNs) to outperform commonplace be aware embedding in certain cases
• Identify difficulties for which Recurrent Neural community (RNN) suggestions are suitable
• Explore the method required to enforce Autoencoders
• Evolve a deep neural community utilizing reinforcement learning

About the Author

Antonio Gulli is a software program govt and enterprise chief with a fondness for developing and handling worldwide technological expertise, innovation, and execution. he's a professional in se's, on-line companies, desktop studying, info retrieval, analytics, and cloud computing. to date, he has been fortunate sufficient to achieve specialist adventure in 4 diversified international locations in Europe and controlled humans in six diverse international locations in Europe and the US. Antonio served as CEO, GM, CTO, vice chairman, director, and location lead in a number of fields spanning from publishing (Elsevier) to customer web (Ask.com and Tiscali) and high-tech R&D (Microsoft and Google).

Sujit Pal is a know-how study director at Elsevier Labs, engaged on development clever platforms round study content material and metadata. His fundamental pursuits are info retrieval, ontologies, average language processing, computer studying, and disbursed processing. he's at the moment engaged on photograph class and similarity utilizing deep studying types. ahead of this, he labored within the purchaser healthcare undefined, the place he helped construct ontology-backed semantic seek, contextual advertisements, and EMR info processing structures. He writes approximately expertise on his web publication at Salmon Run.

Table of Contents

1. Neural Networks Foundations
2. Keras set up and API
3. Deep studying with ConvNets
4. Generative hostile Networks and WaveNet
5. Word Embeddings
6. Recurrent Neural community — RNN
7. Additional Deep studying Models
8. AI online game Playing
9. Conclusion