January 17th Wednesday

The FIFTH SV_EXPERIENCE (January 17th Wednesday)

• Title : How to build a chatbot
• Date and Time : Wednesday, 17 Jan 2018, 18:00 am ~ 21:00 pm
• Place : IBM
• Attendants : Alphateam ( GaeulGo, Minjun Koo, Hwancheol Kang, Gihyeon Yang )

Contents

• Agenda
  1. Open Discussion
  2. Why Chatbots
  3. Cuts of a chatbot
  4. Hands-On
  5. Follow Up : Topics / Homework

2. Why Chatbots?

• It’s because automatic solutions are simply more scalable. Besides, its cheaper, more efficient, and often more accrurate than a human.

3. Cuts of a chabot

• Intents / Entites / Dialog
• It’s good to be mind these 3 things.
• Intents : An intent is the goal or purpost of the user’s input
• Entities : An entity is a portion of the user’s input that you can use to provide a dfferent response to a particular intent.
• Dialog

4. Hands-On

• register for ibm cloud : ibm.biz/ibmcloudchatbot -> my gate way to fall along the lab
• try our code pattern : ibm.biz/chatbot011718 -> the closest lab experience the lab

• We got a paper written about ‘Watson Conversation’. There are several steps to start Chatbot.
• The first step is ‘Desigining your bot’. We have to name the chatbot and what they do. I wanted to apply our project, so I chose ‘alphabot’ as the chatbot name. And its role is ‘to find the local cafe’.
• The second step is ‘Train Watson Conversation Service’. I signed in ‘IBM Cloud’ and I launched the ‘watson conversation’.
• I entered the name and create chatbot. Continuously, I created the intent, entity, and dialog. It was hard to understand at first but it gets more interesting as time goes by. Finally, I made my first chatbot, ‘Alphabot’, and it took only 1 hour.
• We made a chatbot called ‘alpha_search’. It was easy to make it because the UI of IBM’s conversation is made easy to follow.

Talk our Feelings

• Gaeul Go : It was amazed that I can make a chatbot in 1 hour. The level and speed of the lecture was suitable for us. It was an excited experience.
• Minjun Koo : I thought it would be a lot harder because the host is IBM, but I really appreciated it teaching in detail from the beginning. Gaeul leader who picks up the meet-up well is super cool.
• Gihyeon Yang : I have heard a lot of information about the technology actually used in this MVP BOOTCAMP. Of course, we used a platform different from the chatbot platform we used, but the principles and concepts overlap easily.
• Hwancheol Kang : It was uncomfortable because Wi-Fi was not available, but the staff worked well. I felt that the IBM’s Watson platform was cleaner and more intuitive than the dialogflow in MVP. Both are the same chatbots, so their usage is similar. I had a lot of questions while dealing with chatbots, and I was able to solve some of them.