Machine Learning Behind Chatbots

Google Assistant and Siri have tickled our curiosity enough to dive deep and understand under the hood technologies that make a chatbot.
Though we don’t have Google level of data to create a generalized chatbot, we can use the existing NLP engines and create chatbots that produce valuable results in a specific domain.

This talk is structured from my quest to build a chatbot engine and what I understood by demystifying an open-source chatbot engine.

Transcript

1. Machine Learning Behind Chatbots Bhavani Ravi @geeky_bhavani
2. About Me - Software Engineer Orangescape - Chatbots & Python specialist  - Build2learn, Women Techmakers - I live in twitter - @geeky_bhavani
3. What is a Chatbot?
4. Domain Specific Chatbots
6. Architecture
8. Terminologies
9. Domain Specific Chatbots
10. How to build your First Chatbot?
11. Tutorials Tell You Alot LUIS
12. What’s Inside the Bot Engine?
13. Natural Language Processing 1. Intent Classification 2. Entity Extraction
14. Intent Classification
15. Training Data
16. SVM Classifier Suggest me some italian restaurant search_restuarant
17. Preprocessing
18. Text To Words
19. Feature Generation - Word2Vector
20. Label Encoding Label Encoded Value Greet 0 restaurant_search 1 bye 2
21. Preprocessed Data
22. Training the Classifier
23. Tokenize Suggest me some italian restaurant SVM model Featurize Classify Label Encode Training Pipeline
24. Tokenize Hi Greet Featurize Decode Label Classify Deployment Pipeline
25. Entity Extraction NER using Conditional Random Field
26. Named Entity Recognition Suggest me some Chinese restaurant Cuisine: Chinese
27. Training Data
28. Tokenize
29. Feature Extraction
30. Labels - BILOU tags
31. Prediction
32. Tokenize Suggest me some italian restaurant Cuisine: Italian Featurize Predict Label Encode BILOU tagging Entity Extraction Pipeline
33. Tokenize Suggest me some italian restaurant Cuisine: Italian Featurize Label Decode Predict Deployment
35. Some Interesting Facts 1. We never remove stop words 2. NER is used in POS Tagging 3. It is so easy to understand the ML behind chatbot 4. None of this is important for you to get started with building chatbots
36. Questions? Checkout the whole blog series medium.com/@bhavaniravi/series Bhavani Ravi @geeky_bhavani