Conversational AI: Building clever chatbots
summary
Open source Rasa is the concrete product: a pipeline using scikit‑learn, SpaCy and character n‑grams with a conditional random field for NER, SVM or simple neural nets for intent, and a recurrent model for dialogue management. A demo shows intent/NER working from only 40 training examples and an online learning loop that corrects wrong actions and retrains the dialogue model; reinforcement learning is explored but not yet productionized. Watch if you want reproducible code, exact models, and working demos.
AI-written summary · about
chapters (16)
- 0:00Conversational AI intro
- 1:48Why conversational AI matters
- 3:32Open source chatbot frameworks
- 3:54Natural language understanding fundamentals
- 7:49Demo: intent/entity extraction
- 9:58Handcrafted routes fail
- 11:02Data-driven dialogue management
- 14:11Training data generation
- 15:36Live demo: training loop
- 19:12Open challenges in AI
- 21:45End-to-end chatbot networks
- 22:49Small data techniques
- 23:51Q&A: Intent weight
- 25:37Q&A: Intent classification
- 27:31Q&A: RL for dialogue
- 28:56Q&A: Monetization
related talks
01h 38mConversational AI with Rasa Core & NLU
31mRasa: open source conversational AI to build next generation chatbots
01h 27mDeprecating the state machine: building conversational AI with the Rasa stack
01h 31mGoing beyond “Sorry, I didn’t get that” building AI assistants that scale...
01h 29mGoing Beyond “Sorry, I Didn’t Get That”: Building AI Assistants that Scale...
33mBuilding the next-generation Conversational AI with Python and Deep Learning
41mConversational Agents at Scale: Retrieval and Generative approaches
38mAI-Powered Conversational Interfaces