Book contents
- Frontmatter
- Contents
- List of Figures
- List of Tables
- Preface
- 1 Introduction
- 2 The Perceptron
- 3 Logistic Regression
- 4 Implementing Text Classification Using Perceptron and Logistic Regression
- 5 Feed-Forward Neural Networks
- 6 Best Practices in Deep Learning
- 7 Implementing Text Classification with Feed-Forward Networks
- 8 Distributional Hypothesis and Representation Learning
- 9 Implementing Text Classification Using Word Embeddings
- 10 Recurrent Neural Networks
- 11 Implementing Part-of-Speech Tagging Using Recurrent Neural Networks
- 12 Contextualized Embeddings and Transformer Networks
- 13 Using Transformers with the Hugging Face Library
- 14 Encoder-Decoder Methods
- 15 Implementing Encoder-Decoder Methods
- 16 Neural Architectures for Natural Language Processing Applications
- Appendix A Overview of the Python Language and Key Libraries
- Appendix B Character Encodings: ASCII and Unicode
- References
- Index
6 - Best Practices in Deep Learning
Published online by Cambridge University Press: 01 February 2024
- Frontmatter
- Contents
- List of Figures
- List of Tables
- Preface
- 1 Introduction
- 2 The Perceptron
- 3 Logistic Regression
- 4 Implementing Text Classification Using Perceptron and Logistic Regression
- 5 Feed-Forward Neural Networks
- 6 Best Practices in Deep Learning
- 7 Implementing Text Classification with Feed-Forward Networks
- 8 Distributional Hypothesis and Representation Learning
- 9 Implementing Text Classification Using Word Embeddings
- 10 Recurrent Neural Networks
- 11 Implementing Part-of-Speech Tagging Using Recurrent Neural Networks
- 12 Contextualized Embeddings and Transformer Networks
- 13 Using Transformers with the Hugging Face Library
- 14 Encoder-Decoder Methods
- 15 Implementing Encoder-Decoder Methods
- 16 Neural Architectures for Natural Language Processing Applications
- Appendix A Overview of the Python Language and Key Libraries
- Appendix B Character Encodings: ASCII and Unicode
- References
- Index
Summary
The previous chapter introduced feed-forward neural networks and demonstrated that, theoretically, implementing the training procedure for an arbitrary feed-forward neural network is relatively simple. Unfortunately, neural networks trained this way will suffer from several problems such as stability of the training process – that is, slow convergence due to parameters jumping around a good minimum – and overfitting. In this chapter, we will describe several practical solutions that mitigate these problems. In particular, we discuss minibatching, multiple optimization algorithms, other activation and cost functions, regularization, dropout, temporal averaging, and parameter initialization and normalization.
Keywords
- Type
- Chapter
- Information
- Deep Learning for Natural Language ProcessingA Gentle Introduction, pp. 87 - 106Publisher: Cambridge University PressPrint publication year: 2024