X Международная студенческая научная конференция Студенческий научный форум - 2018

Nowadays many computer systems are able to analyze a language and in connection with rapid development of information technologies and continuous increase of volumes of information, available in a global network, the questions of effective information searching become more and more important. But standard scanning with the help of key words doesn’t give desired results. Because , first of all, 80 percent of information in the Internet is unstructured. Secondly, the linguistic and the semantic relationships between key words are not taken into account.

That’s why Natural Language Processing (NLP) technologies and based on them Question –Answering Systems (QAS) are actively developed.

Question–Answering Systems are complex informational systems which are able to answer natural language questions ( questions in natural language) In other words it allows a man to conduct with a machine a dialog in a human language. So as Rohini Srihari and Wei Li has concluded (1999), Natural language QA “is an ideal test bed for demonstrating the power of IE (Information Extraction)…there is a natural co-operation between IE and IR (Information Retrieval)”. So Question-Answering (QA) is a system based on information extraction, information retrieval and natural language processing (NLP).

The first Question –Answering Systems were appeared in 60^th and among them were BASEBALL and LUNAR systems. Both systems were sufficiently effective realized but they were closed-domain and oriented to the concrete subject domain. For example, BASEBALL was used for a dialogue about results of baseball league competitions of the USA, and LUNAR system deals with space.

But now one of the most famous Question –Answering System is IBM Watson system built by the IBM researchers under the direction of David Ferrucci. In 2011 IBM Watson defeated two human champions in the TV quiz show Jeopardy in America.

We try to understand what advantages IBM Watson system has and how it works. That is the major topic of this paper.

In February 2011 there was an incredible event not only in the world of information technology. Among the participants of TV show Jeopardy! was a super computer and more than that it became the winner. Jeopardy! is a very popular in many countries around the world TV game-quiz. In compare chess match Deep Blue against Garry Kasparov, triumph of IBM Watson system in Jeopardy! had a huge value both for the future in general and for development of the analytical systems in particular.

The last decade has seen great advances and interest in the area of IE. In the US, the DARPA sponsored Tipster Text Program [Grishman 1997] and the Message Understanding Conferences (MUC) [MUC-7 1998] have been the driving force for developing this technology. The goals of IBM Research are to advance computer science by exploring new ways for computer technology to affectscience, business, and society. Watson is named after IBM’s founder, Thomas J. Watson.

After three years ofintense research and development by a coreteam of about 20 researchers, Watson is performing at human expert levels in terms of precision,confidence, and speed at the Jeopardyquiz show. The results strongly suggest that DeepQA is an effective and extensible architecturethat can be used as a foundation for combining,deploying, evaluating, and advancing awide range of algorithmic techniques to rapidlyadvance the field of question answering (QA).

DeepQA is aneffective and extensible architecture that may beused as a foundation for combining, deploying, evaluating, and advancing a wide range of algorithmictechniques to rapidly advance the field of QA.

Question-Answering system (QA) are designed to find precise answers to the questions posed in natural language (NLP). Now Watson is able to find answers to 85% of the questions within 5 seconds. At the stage of decomposition issue is broken into pieces. Then the process of generating hypotheses begins. The knowledge base is conducted primary search for pieces of text that contain information about the parts of the question. In this step, Watson throws a few hundred hypotheses.

In the next step the hypothesis are filtered using algorithms. After filtering remains approximately 100 hypotheses.

Then for each hypothesis takes place search additional facts in the knowledge base . Stage evaluation of hypotheses is the most complex computational point of view, because the score of each hypothesis consists of more than 50 components. On stage synthesis candidate responses are combined into one response.

The last stage is the final score is calculated estimation of reliability of hypothesis and of each candidate's reply.

Now Watson is able to find answers to 85% of the questions within 5 seconds.Watson could find a wider application: in health care, law, finance — wherever there is a well defined knowledge base (encyclopaedias, manuals, legislation, etc.) used to obtain the answers.

Список литературы

[Электронный рессурс]

https://www.youtube.com/watch?v=3G2H3DZ8rNc

https://www.youtube.com/watch?v=qAocQF2xXGg