Artificial Intelligence : Using Watson API in RGTC

Artificial Intelligence, better known as AI, is gaining ground to becoming one of the trends the modern age. When we think of AI, it is more associated with robots. This blog describes what Rococo Global Technologies Corporation (RGTC) has already done with AI(IBM Watson in particular).

So what exactly is AI? As defined in Wikipedia:

Artificial intelligence (AI) is intelligence exhibited by machines. In computer science, an ideal “intelligent” machine is a flexible rational agent that perceives its environment and takes actions that maximize its chance of success at some goal. Colloquially, the term “artificial intelligence” is applied when a machine mimics “cognitive” functions that humans associate with other human minds, such as “learning” and “problem solving”.

As a result, machines today are exhibiting more human-like behavior.

Recently, there has been a rise of robots being integrated with AIs. In addition, the AIs perform cognitive tasks and to analyze huge amounts of data. One such robot being integrated with AI is Pepper. For more background on Pepper and its functions, visit this link. Below are some of the APIs used by Pepper to integrate with Watson:

Conversation

IBM Watson Conversation logo

The Conversation API allows an application to use its trained language capabilities to converse with users.
The AI itself isn’t ready made. The developers need to train the AI to be able to give the users the nearest response to their questions. To create the application, all you need is four steps(taken from Conversation documentation):

1. Train Watson to understand your users’ input with example utterances: Intents and Examples.
2. Identify the terms that may vary in your users’ input: Entities.
3. Create the responses to your user’s questions: Dialog Builder.
4. Test and Improve.

To test the application, the user can just ask a question to the application and it will return a response. The response comes in a form of a Javascript Object Notation(JSON) format. This file contains the information of the response such as its reply and the topic of the conversation.


Visual Recognition

IBM Watson Visual Recognition logo

This is another cognitive function that recognizes and classifies images. The classification involves scenes, objects, faces, text and other content. The flow of the application is described below.

Artificial Intelligence IBM Visual Recognition process flow

Based on the flow above, the application is fairly easy to use . First, feed the image to the application. After that, it will return the details of the image. Unlike the Conversation API which has to be trained, the Visual Recognition API already has a set of trained responses for a particular image. If you want to add custom classifiers, you can train new classifiers to the API. The return values are also in JSON format.


Working with Artificial Intelligence in RGTC

Watson, at first glance, is fairly easy to use. It has a lot of tutorials and demos on how to use the API. In addition, its documentation is easy to understand. However, since Watson APIs are still new, there is only a handful of forums to ask questions.

The experience of working with AI requires creative and logical thinking. It takes some time to learn these things but RGTC has coped up very well. In the start, RGTC has started developing prototypes using the APIs. After the prototypes were stable enough, RGTC tried to integrate the APIs using Pepper as a medium. As a result, Pepper now uses Watson for other cognitive functions aside from using its own AI.

Even with these new technologies, RGTC has already made applications that use the above-mentioned APIs with the use of Pepper. There are two effects for this. First, it is now possible for Pepper to use cognitive functions other than using his own AI. Second, it provides opportunities for RGTC to market its knowledge with robots and AI to have more of the same projects.

RGTC will continue to research other functions of Watson to integrate with Pepper. The aim now is to fully utilize Watson’s capabilities alongside Pepper’s AI. In addition, Watson integration will enhance Pepper’s ability to interact with humans.

Conclusion

Artificial Intelligence has come a long way since its birth. As a matter of fact, cognitive functions being created are becoming more complex. The goal of creating an excellent AI is not far away. Pepper with Watson is a step towards achieving a perfect robot with a perfect AI.