The recent decades have seen phenomenal advances in information Technology and its applications in diverse fields, including those which seemingly require human intelligence. This underlines the need for powerful human-machine interaction (HMI) systems. People learn (human-human) interaction skills spontaneously and gradually by observation and by actual interaction with parents, family and other children as well as in larger social settings. Given that human-machine interaction is the need of the hour, a need spurred by technological changes in the society, the importance of developing powerful and easy to use HMI systems cannot be overemphasized.
In this monograph we specifically focus on HMI with particular emphasis on making them user friendly and, as an ideal goal, as natural as human-human interaction.
In "Friendly Interfaces between Humans and Machine" we specifically focus on HMI with particular emphasis on making them user friendly and, as an ideal goal, as natural as human-human interaction.
For realizing this, we need firstly to understand the expected behavior of the person interacting with the machine. We need secondly to recognize and appreciate technological limitations that constrain implementation of systems that make human-machine interaction both possible and usable.
At present, on the one hand researchers keep focusing on bettering the technology without necessarily worrying adequately about its usability; on the other hand, implementers and developers remain primarily interested in building quick solutions, without worrying about the capabilities and limitations of technology.
In this monograph we try to take a middle path:
We hope that our approach will enable developers to build human-machine interfaces that are easy to implement and use. Hopefully, it will also help the beginner student to obtain a clear understanding of the subject in proper perspective.
To start with, we review the evolution of language as a spontaneous, natural phenomenon in the overall scheme of the evolutionary development of living beings. Such a study is necessary and will be very useful in the context of attempting to use natural language as a convenient medium of communication between humans on the one hand and machines on the other.
We then examine possible approaches to understand and represent the meaning and the common aspects of human-human and human-machine interactions. We introduce our key concept-key word (minimal parsing) approach as a convenient and realistic way to implement usable HMI systems. We also describe a working question answering (QA) system along with a number of practical applications based on our minimal parsing approach.