VisualDialogAI

• 

  food

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

Inspiration

Visual Dialog is a novel task that requires an AI agent to hold a meaningful dialog with humans in natural, conversational language about visual content.

What it does

Specifically, given an image, a dialog history, and a follow-up question about the image, the agent has to answer the question.

How I built it

I previously built it using Convolutional Neural Networks but now I used IBM Watson's deep learning engine to build it

Challenges I ran into

The student account doesn't allow me to work with alot of APIs. The Allowed API's are also limited

Accomplishments that I'm proud of

I worked with the leading AI reasoning engines "IBM Watson" and build my web app

What I learned

Its the first time I have worked with APIs and I learned web development and API interfaces

What's next for VisualDialogAI

Next is to build nested reasoning layers so to answer questions that are related to one other

Built With

• ibm-bluemix-conversation
• ibm-bluemix-conversation-api
• ibm-bluemix-texttospeecapi
• r