kTutor

• 

  kTutor mind-map

Inspiration

I really didn't have any ideas before joining this Hackathon. I didn't really understand 5G, and wanted to learn more about it. However, after sitting through the Samsung and IBM workshops, I was inspired by a couple of ideas. I wanted to take the challenge of developing this app myself since this is my first hackathon as a Front-End Developer. Flying solo, has allowed me to learn faster and improve my organizational skills.

In addition, I had been taking Korean language courses, but nothing really stuck until I immersed myself in Korean media (ie K-Dramas). I have been hooked on a drama titled "Home for Summer," which uses uncommon vocabulary and phrases.

I wasn't sure if Watson would allow me to use Korean Language, in which case I would have had to translate everything into Spanish and use Spanish Novellas.

What it does

kTutor is a chatbot that teaches Korean language while you watch Korean Dramas. The goal is to enhance learning through fun.

How I built it

I used IBM's Cloud, IBM's Chatbot API and the Watson library for Korean language. I also used Google Translate to look up most of the words and phrases for the curriculum, since I didn't have a Korean keyboard.

Challenges I ran into

Technical issues probably due to my inexperience with coding. Another technical challenge I had is that I had to copy and paste Korean characters from my iPhone via email since I do not have a Korean keyboard. The IBM team has been great (ie Dave Nugent/Upkar Lidder) with helping troubleshoot, and one-on-one tutorials. The last was figuring out the logic of a "Repeat after me" assessment, but not having the time to code it in less than a weekend.

Accomplishments that I'm proud of

I'm proud that I completed the working prototype, and that I discovered a new way to design these types of interactions.

What I learned

A lot... I was really impressed with the IBM presentation on NLP, Image Analysis, Chatbot, Sentiment Analysis, etc. I had always visited the site to research AI generally, but never knew there was a sandbox to try Watson-features. I have an idea now of how to use sentiment analysis and text-to-speech.

What's next for kTutor

Maybe refining the app to use "slots." Maybe explore Watson's NLP features and explore the "Repeat after me" feature, how to embed feedback, comment in the media while the media is played or being broadcasted (an idea that IBM Upkar Lidder and I bounced around).

Built With

• cloud
• ibm
• ibm-watson