- 1. CMU Neural Nets for NLP 2020 (1) - Introduction
- 2. CMU Neural Nets for NLP 2020 (2) - Language Modeling Efficiency_Training Tri
- 3. CMU Neural Nets for NLP 2020 (3) - Convolutional Neural Networks for Text
- 4. CMU Neural Nets for NLP 2020 (4) - Recurrent Neural Networks
- 5. CMU Neural Nets for NLP 2020 (5) - Efficiency Tricks for Neural Nets
- 6. CMU Neural Nets for NLP 2020 (7) - Attention
- 7. CMU Neural Nets for NLP 2020 (8) - Distributional Semantics and Word Vectors
- 8. CMU Neural Nets for NLP 2020 (9) - Sentence and Contextual Word Representatio
- 9. CMU Neural Nets for NLP 2020 (10) - Debugging Neural Nets (for NLP)
- 10. CMU Neural Nets for NLP 2020 (11) - Structured Prediction with Local Indepen
- 11. CMU Neural Nets for NLP 2020 (12) - Generating Trees Incrementally
- 12. CMU Neural Nets for NLP 2020 (13) - Generating Trees Incrementally
- 13. CMU Neural Nets for NLP 2020 (14) - Search-based Structured Prediction
- 14. CMU Neural Nets for NLP 2020 (15) - Minimum Risk Training and Reinforcement
- 15. CMU Neural Nets for NLP 2020 (16) - Advanced Search Algorithms
- 16. CMU Neural Nets for NLP 2020 (17) - Adversarial Methods
- 17. CMU Neural Nets for NLP 2020 (18) - Models w_ Latent Random Variables
- 18. CMU Neural Nets for NLP 2020 (19) - Unsupervised and Semi-supervised Learnin
- 19. CMU Neural Nets for NLP 2020 (20) - Multitask and Multilingual Learning
- 20. CMU Neural Nets for NLP 2020 (21) - Document Level Models
- 21. CMU Neural Nets for NLP 2020 (22) - Neural Nets Knowledge Bases
- 22. CMU Neural Nets for NLP 2020 (23) - Machine Reading w_ Neural Nets
- 23. CMU Neural Nets for NLP 2020 (24) - Natural Language Generation
- 24. CMU Neural Nets for NLP 2020 (25) - Model Interpretation
神經(jīng)網(wǎng)絡(luò)促進了語言建模的快速發(fā)展，并且已被用于優(yōu)化很多其他NLP任務(wù)，甚至解決很多過去不容易的新問題。本課程將首先對神經(jīng)網(wǎng)絡(luò)進行簡要概述，然后主要講解如何將神經(jīng)網(wǎng)絡(luò)應(yīng)用于自然語言問題。每個部分都將以自然語言任務(wù)入手，介紹一個特定的問題或現(xiàn)象，描述為何難以建模，并演示一些旨在解決該問題的模型。在這樣做的過程中，該課程將涵蓋可用于創(chuàng)建神經(jīng)網(wǎng)絡(luò)模型的不同技術(shù)，包括處理大小可變和結(jié)構(gòu)化的句子，有效處理大數(shù)據(jù)，半監(jiān)督和無監(jiān)督學習，結(jié)構(gòu)化預測以及多語言建模。
