Word Embeddings for Natural Language Processing in Python

Slides for my talk on word embeddings at PyCon Italy 2017 (PyCon Otto):
https://www.pycon.it/conference/talks/word-embeddings-for-natural-language-processing-in-python
Abstract:
Word embeddings are a family of Natural Language Processing (NLP) algorithms where words are mapped to vectors in low-dimensional space. The interest around word embeddings has been on the rise in the past few years, because these techniques have been driving important improvements in many NLP applications like text classification, sentiment analysis or machine translation.

In this talk we’ll describe the intuitions behind this family of algorithms, we’ll explore some of the Python tools that allow us to implement modern NLP applications and we’ll conclude with some practical considerations.

Transcript

1. Word Embeddings 
 for NLP in Python Marco Bonzanini
 PyCon Italia 2017
2. Nice to meet you
3. WORD EMBEDDINGS?
4. Word Embeddings Word Vectors Distributed Representations = =
5. Why should you care?
6. Why should you care? Data representation
 is crucial
7. Applications
8. Applications • Classification / tagging • Recommendation Systems • Search Engines (Query Expansion) • Machine Translation
9. One-hot Encoding
10. One-hot Encoding Rome Paris Italy France = [1, 0, 0, 0, 0, 0, …, 0] = [0, 1, 0, 0, 0, 0, …, 0] = [0, 0, 1, 0, 0, 0, …, 0] = [0, 0, 0, 1, 0, 0, …, 0]
11. One-hot Encoding Rome Paris Italy France = [1, 0, 0, 0, 0, 0, …, 0] = [0, 1, 0, 0, 0, 0, …, 0] = [0, 0, 1, 0, 0, 0, …, 0] = [0, 0, 0, 1, 0, 0, …, 0] Rome Paris word V
12. One-hot Encoding Rome Paris Italy France = [1, 0, 0, 0, 0, 0, …, 0] = [0, 1, 0, 0, 0, 0, …, 0] = [0, 0, 1, 0, 0, 0, …, 0] = [0, 0, 0, 1, 0, 0, …, 0] V = vocabulary size (huge)
13. Bag-of-words
14. Bag-of-words doc_1 doc_2 … doc_N = [32, 14, 1, 0, …, 6] = [ 2, 12, 0, 28, …, 12] … = [13, 0, 6, 2, …, 0]
15. Bag-of-words doc_1 doc_2 … doc_N = [32, 14, 1, 0, …, 6] = [ 2, 12, 0, 28, …, 12] … = [13, 0, 6, 2, …, 0] Rome Paris word V
16. Word Embeddings
17. Word Embeddings Rome Paris Italy France = [0.91, 0.83, 0.17, …, 0.41] = [0.92, 0.82, 0.17, …, 0.98] = [0.32, 0.77, 0.67, …, 0.42] = [0.33, 0.78, 0.66, …, 0.97]
18. Word Embeddings Rome Paris Italy France = [0.91, 0.83, 0.17, …, 0.41] = [0.92, 0.82, 0.17, …, 0.98] = [0.32, 0.77, 0.67, …, 0.42] = [0.33, 0.78, 0.66, …, 0.97] n. dimensions << vocabulary size
19. Word Embeddings Rome Paris Italy France = [0.91, 0.83, 0.17, …, 0.41] = [0.92, 0.82, 0.17, …, 0.98] = [0.32, 0.77, 0.67, …, 0.42] = [0.33, 0.78, 0.66, …, 0.97]
20. Word Embeddings Rome Paris Italy France = [0.91, 0.83, 0.17, …, 0.41] = [0.92, 0.82, 0.17, …, 0.98] = [0.32, 0.77, 0.67, …, 0.42] = [0.33, 0.78, 0.66, …, 0.97]
21. Word Embeddings Rome Paris Italy France = [0.91, 0.83, 0.17, …, 0.41] = [0.92, 0.82, 0.17, …, 0.98] = [0.32, 0.77, 0.67, …, 0.42] = [0.33, 0.78, 0.66, …, 0.97]
22. Word Embeddings Rome Paris Italy France
23. Word Embeddings Paris + Italy - France ≈ Rome Rome
24. THE MAIN INTUITION
25. Distributional Hypothesis
26. –J.R. Firth 1957 “You shall know a word 
 by the company it keeps.”
27. –Z. Harris 1954 “Words that occur in similar context tend to have similar meaning.”
28. Context ≈ Meaning
29. I enjoyed eating some pizza at the restaurant
30. I enjoyed eating some pizza at the restaurant Word
31. I enjoyed eating some pizza at the restaurant The company it keeps Word
32. I enjoyed eating some pizza at the restaurant I enjoyed eating some fiorentina at the restaurant
33. I enjoyed eating some pizza at the restaurant I enjoyed eating some fiorentina at the restaurant
34. I enjoyed eating some pizza at the restaurant I enjoyed eating some fiorentina at the restaurant Same context
35. I enjoyed eating some pizza at the restaurant I enjoyed eating some fiorentina at the restaurant Same context Pizza = Fiorentina ?
36. A BIT OF THEORY word2vec
39. Vector Calculation
40. Vector Calculation Goal: learn vec(word)
41. Vector Calculation Goal: learn vec(word) 1. Choose objective function
42. Vector Calculation Goal: learn vec(word) 1. Choose objective function 2. Init: random vectors
43. Vector Calculation Goal: learn vec(word) 1. Choose objective function 2. Init: random vectors 3. Run gradient descent
44. I enjoyed eating some pizza at the restaurant
45. I enjoyed eating some pizza at the restaurant
46. I enjoyed eating some pizza at the restaurant
47. I enjoyed eating some pizza at the restaurant Maximise the likelihood 
 of the context given the focus word
48. I enjoyed eating some pizza at the restaurant Maximise the likelihood 
 of the context given the focus word P(i | pizza) P(enjoyed | pizza) … P(restaurant | pizza)
49. Example I enjoyed eating some pizza at the restaurant
50. I enjoyed eating some pizza at the restaurant Iterate over context words Example
51. I enjoyed eating some pizza at the restaurant bump P( i | pizza ) Example
52. I enjoyed eating some pizza at the restaurant bump P( enjoyed | pizza ) Example
53. I enjoyed eating some pizza at the restaurant bump P( eating | pizza ) Example
54. I enjoyed eating some pizza at the restaurant bump P( some | pizza ) Example
55. I enjoyed eating some pizza at the restaurant bump P( at | pizza ) Example
56. I enjoyed eating some pizza at the restaurant bump P( the | pizza ) Example
57. I enjoyed eating some pizza at the restaurant bump P( restaurant | pizza ) Example
58. I enjoyed eating some pizza at the restaurant Move to next focus word and repeat Example
59. I enjoyed eating some pizza at the restaurant bump P( i | at ) Example
60. I enjoyed eating some pizza at the restaurant bump P( enjoyed | at ) Example
61. I enjoyed eating some pizza at the restaurant … you get the picture Example
62. P( eating | pizza )
63. P( eating | pizza ) ??
64. P( eating | pizza ) Input word Output word
65. P( eating | pizza ) Input word Output word P( vec(eating) | vec(pizza) )
66. P( vout | vin ) P( vec(eating) | vec(pizza) ) P( eating | pizza ) Input word Output word
67. P( vout | vin ) P( vec(eating) | vec(pizza) ) P( eating | pizza ) Input word Output word ???
68. P( vout | vin )
69. cosine( vout, vin )
70. cosine( vout, vin ) [-1, 1]
71. softmax(cosine( vout, vin ))
72. softmax(cosine( vout, vin )) [0, 1]
73. softmax(cosine( vout, vin )) P (vout | vin) = exp(cosine(vout , vin)) P k 2 V exp(cosine(vk , vin))
74. Vector Calculation Recap
75. Vector Calculation Recap Learn vec(word)
76. Vector Calculation Recap Learn vec(word) by gradient descent
77. Vector Calculation Recap Learn vec(word) by gradient descent on the softmax probability
78. Plot Twist
81. Paragraph Vector a.k.a. doc2vec i.e. P(vout | vin, label)
82. A BIT OF PRACTICE
84. pip install gensim
85. Case Study 1: Skills and CVs
86. Case Study 1: Skills and CVs Data set of ~300k resumes Each experience is a “sentence” Each experience has 3-15 skills Approx 15k unique skills
87. Case Study 1: Skills and CVs from gensim.models import Word2Vec fname = 'candidates.jsonl' corpus = ResumesCorpus(fname) model = Word2Vec(corpus)
88. Case Study 1: Skills and CVs model.most_similar('chef') [('cook', 0.94), ('bartender', 0.91), ('waitress', 0.89), ('restaurant', 0.76), ...]
89. Case Study 1: Skills and CVs model.most_similar('chef',
 negative=['food']) [('puppet', 0.93), ('devops', 0.92), ('ansible', 0.79), ('salt', 0.77), ...]
90. Case Study 1: Skills and CVs Useful for: Data exploration Query expansion/suggestion Recommendations
91. Case Study 2: Beer!
92. Case Study 2: Beer! Data set of ~2.9M beer reviews 89 different beer styles 635k unique tokens 185M total tokens
93. Case Study 2: Beer! from gensim.models import Doc2Vec fname = 'ratebeer_data.csv' corpus = RateBeerCorpus(fname) model = Doc2Vec(corpus)
94. Case Study 2: Beer! from gensim.models import Doc2Vec fname = 'ratebeer_data.csv' corpus = RateBeerCorpus(fname) model = Doc2Vec(corpus) 3.5h on my laptop … remember to pickle
95. Case Study 2: Beer! model.docvecs.most_similar('Stout') [('Sweet Stout', 0.9877), ('Porter', 0.9620), ('Foreign Stout', 0.9595), ('Dry Stout', 0.9561), ('Imperial/Strong Porter', 0.9028), ...]
96. Case Study 2: Beer! model.most_similar([model.docvecs['Stout']]) 
 [('coffee', 0.6342), ('espresso', 0.5931), ('charcoal', 0.5904), ('char', 0.5631), ('bean', 0.5624), ...]
97. Case Study 2: Beer! model.most_similar([model.docvecs['Wheat Ale']]) 
 [('lemon', 0.6103), ('lemony', 0.5909), ('wheaty', 0.5873), ('germ', 0.5684), ('lemongrass', 0.5653), ('wheat', 0.5649), ('lime', 0.55636), ('verbena', 0.5491), ('coriander', 0.5341), ('zesty', 0.5182)]
98. PCA
99. Dark beers
100. Strong beers
101. Sour beers
102. Lagers
103. Wheat beers
104. Case Study 2: Beer! Useful for: Understanding the language of beer enthusiasts Planning your next pint Classification
105. FINAL REMARKS
106. But we’ve been
 doing this for X years
107. But we’ve been
 doing this for X years • Approaches based on co-occurrences are not new • Think SVD / LSA / LDA • … but they are usually outperformed by word2vec • … and don’t scale as well as word2vec
108. Efficiency
109. Efficiency • There is no co-occurrence matrix
 (vectors are learned directly) • Softmax has complexity O(V)
 Hierarchical Softmax only O(log(V))
110. Garbage in, garbage out
111. Garbage in, garbage out • Pre-trained vectors are useful • … until they’re not • The business domain is important • The pre-processing steps are important • > 100K words? Maybe train your own model • > 1M words? Yep, train your own model
112. Summary
113. Summary • Word Embeddings are magic! • Big victory of unsupervised learning • Gensim makes your life easy
114. Credits & Readings
115. Credits & Readings Credits • Lev Konstantinovskiy (@gensim_py) • Chris E. Moody (@chrisemoody) see videos on lda2vec Readings • Deep Learning for NLP (R. Socher) http://cs224d.stanford.edu/ • “word2vec parameter learning explained” by Xin Rong More readings • “GloVe: global vectors for word representation” by Pennington et al. • “Dependency based word embeddings” and “Neural word embeddings as implicit matrix factorization” by O. Levy and Y. Goldberg
116. THANK YOU @MarcoBonzanini GitHub.com/bonzanini marcobonzanini.com