Automating a Customer Service with Chatbots and Neural Networks
How artificial intelligence and chatbots can drive the future? What are the hidden errors when training a neural network and how to cope with them? The answers to these questions were provided at the recent TensorFlow meetup in Munich.
The time for chatbots is now
Xaver Lehmann of e-bot7 contemplated why artificial intelligence and chatbots in particular may be driving the future. In his session, he talked about market trends and according to Xaver, messaging is a top one now. Exploring the world of chatbots, Xaver explained how they work and how businesses and regular customers can benefit. He exemplified two types of chatbots developed using:
- keyword-based natural language processing
- sophisticated natural language processing with machine learning
After demonstrating a sample technical model of a chatbot, he overviewed the e-bot7 platform developed by the company he works for. The solution aims at enabling businesses to build, integrate, and manage chatbots that would automate commerce and customer services via Facebook Messenger, WhatsApp, Telegram, Kik, and Web Messaging.
Watch the video below for more details.
Hands-on training of a neural network
Starting with a brief overview of TensorFlow and outlining some types of neural networks, Michael Jancen-Widmer of Fiducia & GAD IT AG moved to a practical part of his session and demonstrated how to train a neural network with TensorFlow and TFLearn. In the course of the training, he also explained the attendees how to:
- deal with hidden errors
- find a local minimum with the gradient descent algorithm
For more details, watch the video of Michael’s session.
You can also take a look at his full presentation below.
Join our group to stay tuned with the upcoming events.
- Building a Chatbot with TensorFlow and Keras
- Under-the-Hood Mechanisms of Neural Networks with TensorFlow
- How TensorFlow Can Help to Perform Natural Language Processing Tasks
- Performance of DL Frameworks: Caffe, Deeplearning4j, TensorFlow, Theano, and Torch
About the experts