Intelligent Software understands language and human behavior and provides functions for users in the most convenient way. POSTECH intelligent software Lab (iSOFT) has investigated Spoken Dialog System and Natural Language Processing to build a bridge between humans and computers through language. Educational games for English immersion and automated evaluation and feedback are other challenging research areas for us.

Spoken Dialogue Interface, which permits spoken interaction between users and computers, has much advantage over the previous devices, such as a keyboard, a mouse and a touchscreen. However, Spoken Dialog Processing Technology is usually considered as a high risk and high return: it can’t elicit users’ positive responses until it reaches a certain level, while it is attractive enough in case of success. It needs time and effort to obtain reliability of customers and to satisfy user’s various appetites through Spoken Dialog System. However, it is expected to have a great ripple effect if it could be fully developed and actualized. For this reason, the most valuable enterprise tech companies in the world, such as Apple, Google, Microsoft, and IBM, have encouraged investment in Spoken Dialog and Natural Language Processing research. However, it is domestic researchers who have responsibility on the development of Korean Spoken/ Natural Language Interface. A lack of investment can be an obstacle to the development of processing technology for Korean Language. To compete and survive in world IT markets, the consistent and extensive investment in this field is urgently required.

Spoken Dialog Interface Will Replace Touch Interface
The recently developed Smart Appliances including Smart phones, Smart TVs, etc. appeal to the public mind. Featuring an advanced touch interface, these smart devices provide customers with various user-friendly functions. The convenience of this technology, however, is destined not to be enough sooner or later. We suggest Spoken Dialog Interface as an alternative. The current Spoken Interface technology, such as Google Voice Search, still has a long way to go: what a user speaks is mechanically converted to plain texts. Spoken Dialog System is an advanced technology that analyzes the semantic information, keeps the history and state of the dialog and manages the general flow of the conversation. It is Apple Siri that takes an initiating role in this technology. POSTECH iSOFT Lab has one of the best Spoken Dialog System Technology programs in the world. Spoken Dialog System has a limited ability in that it processes only short sentences, giving and receiving only short messages. POSTECH iSOFT goes beyond the current limitations and covers a broad spectrum of research areas. It investigates possible ways to process more complicated, and extended dialog, to manage topic change dialog, to recognize speaker intention by interpreting human gesture, and so on. Initial versions of software demonstration by POSTECH iSOFT are on You Tube.

POSTECH Immersive English Study
Many researchers in the development of dialog systems incline to the view that the system facilitates human-machine interaction by processing the required action and information appropriately. POSTEC iSOFT Lab, however, seeks to adapt the dialog system to a new domain: English Immersion Education. We’re on the right track with game development for English Immersion Education. It is based on the dialog system for the first time in the world and a large-scale experiment is planned for its effectiveness. POSTECH Immersive English Study (POMY) is a computer program, which provides English learning situation through natural conversation with virtual characters playing games. The application of the dialog system makes the virtual characters of POMY active conversational partners: they respond appropriately and naturally, correct grammatical errors or lead conversation under close guidance for contextual meaning, and give challenging tasks. Previous programs have mechanically passive characters that produce preset responses. We take pride in this program, saying “‘studying by playing”, a long cherished hope of students, can come true through POMY.

Conversational System Q/A
The previous search engine examines its index and provides a listing of the best-matching web pages according to its criteria, when a user enters a self-created query into a search engine typically by using keywords. Time consuming effort in searching and accessing to the information is inevitable. The most popular search engines, such as IBM’s Watson, Apple’s Siri, Google’s Knowledge Graph, and so on, have devoted themselves to the establishment of QA (Question Answering) systems that overcome the limitation of the previous search engines. QA system is a computer science discipline within the fields of information retrieval and natural language processing (NLP), which is concerned with building systems that automatically answer questions posed by humans in a natural language. By pulling answers from an unstructured collection of natural language documents, QA systems build user-friendly search environment. POSTECH iSOFT Lab accelerates the development of QA system with skills in Natural Language Processing. The first step of the development is to build a system to understand and parse questions posed by humans in a natural language. The questions are then transformed into a query. The system constructs its answers by querying a structured database of knowledge or information, usually a knowledge base. Among some valid answers, produced as a query result, the system selects an appropriate answer in response to the user’s question. The recent research of POSTECH iSOFT is focused on the establishment of a knowledge base by extracting and processing voluminous external resources such as Wikipedia.
In addition, POSTECH iSOFT develops a conversational system QA to support intelligent conversation between a user and QA system. It is the combination of skills in dialog system and QA system to better facilitate user information needs. We also have a blue print for a humanized QA system to make an interaction with a computer feel as natural as possible. The humanized QA system is to be developed to respond to user’s questions on the basis of their individualized interests and information. POSTECH iSOFT conversational system QA will be a center of the next generation IT.

Beyond Apple’s Siri or Google’s Voice Search
Current accomplishments are not enough. We will make Natural Language Interface more user-friendly and easily accessible by developing and expanding our skills. It’s only a question of time to take a lead beyond Apple’s Siri or Google’s Google Now. Through dynamic cooperation with domestic and foreign companies, POSTECH iSOFT will reinforce competitiveness of domestic software industry by expanding the application of Spoken Dialog Interface to all kinds of platforms, including smart phones, smart TVs, navigations, and robots. Games for English Immersion Education and Automatic Evaluation/Feedback System are expected to change the previous paradigm of English education and to be a driving force behind the next IT generation.