AI has accelerated research in conversational interaction. With the ubiquity of mobile instant messaging, Chatbot use cases are flourishing. How do we build a chatbot powered by modern AI?
What are the lifecycle differences between a chatbot and traditional web/app products? Drawing on practical experience, ML GDE Huan shares insights and a live demo: using Angular + TensorFlow.js to analyze WeChat messages in the browser, in real time.
Agenda
- Intro to Conversational UI (CUI)
- A brief review of AI development
- Wechaty open-source project intro
- TensorFlow.js overview
- Live Demo: Analyze WeChat messages in the browser with Angular + TensorFlow.js
- Q&A
Speaker
Huan (Li Zixuan), Google Developer Expert (Machine Learning)
- TensorFlow.js Evangelist; Angular enthusiast
- Angel investor at PreAngel focusing on AI/Chatbot startups
- Co-author of “Chatbot from 0 to 1” and “Simple TensorFlow 2”
- Creator of the 10k+ stars open-source Conversational SDK Wechaty
GDE (Machine Learning); Angel Investor. PreAngel Partner, Plug and Play Venture Partner (China). Active angel investor for AI startups; serial entrepreneur with strong technical background.
Presentation
Slides
Source Code
GitHub: https://github.com/huan/ng-plus-wechaty
Tutorials
Wechaty Token: apply a token from Puppet Service Providers
- Apply a token: https://github.com/wechaty/puppet-supports/issues/new/choose
- Puppet Service FAQ: https://wechaty.js.org/docs/puppet-services/faq
DIY a Wechaty Puppet Service Token
export WECHATY_TOKEN=insecure_wechaty_puppet_service_token_diy
export WECHATY_PUPPET_SERVER_PORT=48788
export WECHATY_PUPPET=wechaty-puppet-padlocal
export WECHATY_PUPPET_PADLOCAL_TOKEN=YOUR_PADLOCAL_TOKEN_AT_HERE
export WECHATY_LOG=verbose
docker run \
--rm -ti \
-e WECHATY_LOG \
-e WECHATY_PUPPET \
-e WECHATY_PUPPET_PADLOCAL_TOKEN \
-e WECHATY_PUPPET_SERVER_PORT \
-e WECHATY_TOKEN \
-p "$WECHATY_PUPPET_SERVER_PORT" \
wechaty/wechaty:0.78
Learn more: Puppet Service DIY: https://wechaty.js.org/docs/puppet-services/diy/
Replays
- NG+ 2020:
- GDG Community Talk: