Scholarlink

Transfer learning (TL) is a widely used deep learning approach aimed at minimizing computational resource requirements. This chapter covers (1) a comparison of TL with traditional machine learning (ML); (2) recurrent neural networks (RNNs), a key component of TL, with a focus on core technologies such as long short-term memory (LSTM) networks and bidirectional recurrent neural networks (BRNNs) in natural language processing (NLP) applications; and (3) transformer architecture, including the Bidirectional Encoder Representations from Transformers (BERT) model, as well as related advancements like Transformer-XL and ALBERT models. Additionally, this chapter provides the conceptual foundation for Workshop #6—Transformers with spaCy and TensorFlow, discussed in Chap. 16 .

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Transfer Learning and Transformer Technology

  • 章节

  • 发表日期：

    2025年

查看全文
  • Raymond Lee

摘要

Transfer learning (TL) is a widely used deep learning approach aimed at minimizing computational resource requirements. This chapter covers (1) a comparison of TL with traditional machine learning (ML); (2) recurrent neural networks (RNNs), a key component of TL, with a focus on core technologies such as long short-term memory (LSTM) networks and bidirectional recurrent neural networks (BRNNs) in natural language processing (NLP) applications; and (3) transformer architecture, including the Bidirectional Encoder Representations from Transformers (BERT) model, as well as related advancements like Transformer-XL and ALBERT models. Additionally, this chapter provides the conceptual foundation for Workshop #6—Transformers with spaCy and TensorFlow, discussed in Chap. 16 .