1、 I n t e r n a t i o n a l T e l e c o m m u n i c a t i o n U n i o n ITU-T F.746.3 TELECOMMUNICATION STANDARDIZATION SECTOR OF ITU (11/2015) SERIES F: NON-TELEPHONE TELECOMMUNICATION SERVICES Audiovisual services Intelligent question answering service framework Recommendation ITU-T F.746.3 ITU-T F
2、-SERIES RECOMMENDATIONS NON-TELEPHONE TELECOMMUNICATION SERVICES TELEGRAPH SERVICE Operating methods for the international public telegram service F.1F.19 The gentex network F.20F.29 Message switching F.30F.39 The international telemessage service F.40F.58 The international telex service F.59F.89 St
3、atistics and publications on international telegraph services F.90F.99 Scheduled and leased communication services F.100F.104 Phototelegraph service F.105F.109 MOBILE SERVICE Mobile services and multidestination satellite services F.110F.159 TELEMATIC SERVICES Public facsimile service F.160F.199 Tel
4、etex service F.200F.299 Videotex service F.300F.349 General provisions for telematic services F.350F.399 MESSAGE HANDLING SERVICES F.400F.499 DIRECTORY SERVICES F.500F.549 DOCUMENT COMMUNICATION Document communication F.550F.579 Programming communication interfaces F.580F.599 DATA TRANSMISSION SERVI
5、CES F.600F.699 AUDIOVISUAL SERVICES F.700F.799 ISDN SERVICES F.800F.849 UNIVERSAL PERSONAL TELECOMMUNICATION F.850F.899 HUMAN FACTORS F.900F.999 For further details, please refer to the list of ITU-T Recommendations. Rec. ITU-T F.746.3 (11/2015) i Recommendation ITU-T F.746.3 Intelligent question an
6、swering service framework Summary Recommendation ITU-T F.746.3 defines the intelligent question answering service (IQAS) framework as well as the features, requirements and the functionality to support the intelligent question answering (QA) service. The intelligent QA system is an advanced function
7、 to generate answers to a users questions in a natural language. More and more systems in the future are expected to be equipped with QA functions for an advanced user experience. History Edition Recommendation Approval Study Group Unique ID* 1.0 ITU-T F.746.3 2015-11-29 16 11.1002/1000/12619 Keywor
8、ds Intelligent system, natural language processing, QA, question answering. * To access the Recommendation, type the URL http:/handle.itu.int/ in the address field of your web browser, followed by the Recommendations unique ID. For example, http:/handle.itu.int/11.1002/1000/11830-en. ii Rec. ITU-T F
9、.746.3 (11/2015) FOREWORD The International Telecommunication Union (ITU) is the United Nations specialized agency in the field of telecommunications, information and communication technologies (ICTs). The ITU Telecommunication Standardization Sector (ITU-T) is a permanent organ of ITU. ITU-T is res
10、ponsible for studying technical, operating and tariff questions and issuing Recommendations on them with a view to standardizing telecommunications on a worldwide basis. The World Telecommunication Standardization Assembly (WTSA), which meets every four years, establishes the topics for study by the
11、 ITU-T study groups which, in turn, produce Recommendations on these topics. The approval of ITU-T Recommendations is covered by the procedure laid down in WTSA Resolution 1. In some areas of information technology which fall within ITU-Ts purview, the necessary standards are prepared on a collabora
12、tive basis with ISO and IEC. NOTE In this Recommendation, the expression “Administration“ is used for conciseness to indicate both a telecommunication administration and a recognized operating agency. Compliance with this Recommendation is voluntary. However, the Recommendation may contain certain m
13、andatory provisions (to ensure, e.g., interoperability or applicability) and compliance with the Recommendation is achieved when all of these mandatory provisions are met. The words “shall“ or some other obligatory language such as “must“ and the negative equivalents are used to express requirements
14、. The use of such words does not suggest that compliance with the Recommendation is required of any party. INTELLECTUAL PROPERTY RIGHTSITU draws attention to the possibility that the practice or implementation of this Recommendation may involve the use of a claimed Intellectual Property Right. ITU t
15、akes no position concerning the evidence, validity or applicability of claimed Intellectual Property Rights, whether asserted by ITU members or others outside of the Recommendation development process. As of the date of approval of this Recommendation, ITU had received notice of intellectual propert
16、y, protected by patents, which may be required to implement this Recommendation. However, implementers are cautioned that this may not represent the latest information and are therefore strongly urged to consult the TSB patent database at http:/www.itu.int/ITU-T/ipr/. ITU 2016 All rights reserved. N
17、o part of this publication may be reproduced, by any means whatsoever, without the prior written permission of ITU. Rec. ITU-T F.746.3 (11/2015) iii Table of Contents Page 1 Scope . 1 2 References . 1 3 Definitions 1 3.1 Terms defined elsewhere 1 3.2 Terms defined in this Recommendation . 1 4 Abbrev
18、iations and acronyms 2 5 Conventions 3 6 Introduction . 3 7 Requirements 4 7.1 General requirements for intelligent question answering service and the system to support the QA service . 4 7.2 Natural language processing requirements . 5 7.3 Question analysis requirements 5 7.4 Candidate answer gener
19、ation requirements 6 7.5 Answer inference and generation requirements . 6 8 Functional components and interfaces for intelligent question answering service 7 8.1 Natural language processing . 7 8.2 Question analysis 9 8.3 Candidate answer generation 11 8.4 Answer inference and generation . 12 Biblio
20、graphy. 13 Rec. ITU-T F.746.3 (11/2015) 1 Recommendation ITU-T F.746.3 Intelligent question answering service framework 1 Scope This Recommendation provides an overview of the intelligent question answering service (IQAS) framework and describes the features, general requirements and functionality w
21、hich form a framework to support the intelligent question answering (QA) service. The scope of this Recommendation focuses on QA architecture, devices, servers, interfaces and communication protocols among QA servers and clients in a high-level approach. 2 References None. 3 Definitions 3.1 Terms de
22、fined elsewhere This Recommendation uses the following terms defined elsewhere: 3.1.1 application b-ITU-T Y.101: A structured set of capabilities, which provide value-added functionality supported by one or more services. 3.1.2 content b-ITU-T H.780: A combination of audio, still image, graphic, vid
23、eo, or data. NOTE Variety of formats is classified as the “data“ (e.g., text, encoded values, multimedia description language introduced by b-ITU-T H.760). 3.2 Terms defined in this Recommendation This Recommendation defines the following terms: 3.2.1 co-reference resolution: A function that detects
24、 the preceding referents of the pronouns which replace the noun phrases of the input sentences. 3.2.2 CV_occupation: A named entity tag that belongs to the civilization (CV) category that indicates a persons occupation (e.g., scientist). 3.2.3 CV_prize: A named entity tag that belongs to the civiliz
25、ation (CV) category that indicates a prize (e.g., the Nobel prize). 3.2.4 IsA relation: A hierarchical relation between two words in such a way that the meaning of one word includes that of the other word. 3.2.5 knowledge base: A collection of knowledge resources that consist of structured and unstr
26、uctured data. The knowledge base is used to provide information to the various applications that are related to information provision such as the QA system and search system. 3.2.6 named entity recognition: A function that recognizes named entities such as PLO entities which are people, locations an
27、d organizations from the sentences. The PLO can be decomposed into more specific named entities depending on the applications. 3.2.7 natural language processing: A method that analyses text in natural languages through several processes such as part-of-speech recognition, syntactic analysis and sema
28、ntic analysis. 3.2.8 part-of-speech recognition: A function that recognizes parts of speech (POS) in the sentences and assigns relevant POS tags considering contextual meaning of the target sentences. 2 Rec. ITU-T F.746.3 (11/2015) 3.2.9 question answering: A system that provides answers in natural
29、language to questions which are in natural language form by analysing the questions and all the knowledge resources that are available to the system. 3.2.10 semantic analysis: A function that recognizes the semantic relations among the words around predicates that exist in the same sentence. The sem
30、antic analysis function then generates a semantic predicate-argument structure (PAS). 3.2.11 syntactic analysis: A function that analyses sentence structures and generates dependency relations among words based on dependency grammars. 4 Abbreviations and acronyms This Recommendation uses the followi
31、ng abbreviations and acronyms: ARG Argument CV Civilization DB Database DS Digital Signage ID Identifier IQAS Intelligent Question Answering Service KB Knowledge Base NLP Natural Language Processing NLU Natural Language Understanding NNG Noun General NNP Proper Noun NP_OBJ Noun Phrase Objective NP_S
32、BJ Noun Phrase Subjective PAS Predicate-Argument Structure PLO People, Locations and Organizations POS Parts Of Speech QA Question Answering SL Sub Language SPARQL SPARQL protocol and RDF Query Language SQL Structured Query Language VP_MOD Verb Phrase Modal VV Verb XSNG Auxiliary Suffix from Noun Ge
33、neral XSV Auxiliary Suffix from Verb Rec. ITU-T F.746.3 (11/2015) 3 5 Conventions The following conventions are used in this Recommendation: The keywords “is required to“ indicate a requirement which must be strictly followed and from which no deviation is permitted, if conformance to this Recommend
34、ation is to be claimed. The keywords “is prohibited from“ indicate a requirement which must be strictly followed and from which no deviation is permitted, if conformance to this Recommendation is to be claimed. The keywords “is recommended“ indicate a requirement which is recommended but which is no
35、t absolutely required. Thus, this requirement need not be present to claim conformance. The keywords “is not recommended“ indicate a requirement which is not recommended but which is not specifically prohibited. Thus, conformance with this Recommendation can still be claimed even if this requirement
36、 is present. The keywords “can optionally“ indicate an optional requirement which is permissible, without implying any sense of being recommended. This term is not intended to imply that the vendors implementation must provide the option and the feature can be optionally enabled by the network opera
37、tor/service provider. Rather, it means the vendor may optionally provide the feature and still claim conformance with this Recommendation. 6 Introduction The intelligent question answering (QA) system is an advanced function to generate answers to a users question in a natural language. More and mor
38、e systems in the future are expected to be equipped with QA functions for an advanced user experience. Currently, the most widely recognized QA system is an intelligent agent which is provided in smart phones. Although not a completely intelligent agent able to answer all questions that users may as
39、k, a QA service would be useful for a defined environment. A QA service would be useful if a standardized QA service framework existed which could be applied to various domains such as TV-related services, digital signage (DS) and e-health services. A QA service would also be useful in terms of acce
40、ssibility because natural languages are used to request and provide information conveniently with a speech interface. Figure 1 illustrates the concept of a QA system in the DS domain, where a user requests information about shoe stores from the QA system at a shopping mall. Scenario of an intelligen
41、t question answering service (IQAS) system: An IQAS system at a shopping mall. Michelle visits a shopping mall in Seattle. There is a DS helper which displays information about the various shops at the mall. It also answers customers questions about the shops. Michelle wants to buy a pair of shoes a
42、nd asks the DS helper “Where can I find some nice shoes?“ The DS helper answers “Go to the third floor“. Michelle goes to the third floor where she can find some shoe stores. The answer can also be shown in text form on the screen or can be provided by speech synthesis in a spoken form according to
43、the needs of the user. 4 Rec. ITU-T F.746.3 (11/2015) Figure 1 Digital signage with a QA function: At the mall Figure 2 presents an example of functional blocks in QA architecture. The QA system consists of several functional blocks: natural language processing (NLP), question analysis, candidate an
44、swer generation, answer inference and answer generation. Figure 2 Example of functional blocks in QA architecture 7 Requirements 7.1 General requirements for IQAS and the system to support the QA service The IQAS is required to provide answers to the users various questions considering the users int
45、ention. The IQAS is recommended to provide answers in natural language, whether it is in text or in spoken form depending on the users environment. The IQAS is recommended to provide answers in natural language from multiple information sources. The IQAS is recommended to provide accessibility for d
46、ifferent user needs. Rec. ITU-T F.746.3 (11/2015) 5 The IQAS is recommended to provide answers to the users questions within a proper time span. It is recommended to provide an input editor for the users question whether it consists of a single sentence or multiple sentences. It is recommended that
47、the users question is input in voice or text form depending on the users device. The IQAS is recommended to provide example questions for the users convenience. The IQAS is recommended to provide a manual on the use of the system for the user. The IQAS is recommended to provide a human intervention
48、mechanism where the reliability of the answer is below a predefined threshold value. The IQAS is recommended to be able to save and update profile information such as the history of users questions and answers. The IQAS is recommended to be able to provide answers appropriate to the users level of k
49、nowledge. 7.2 Natural language processing requirements The following requirements relate to NLP in an IQAS and the system to support the IQAS: The IQAS system is recommended to include a NLP module that analyses natural language sources at a deep level. The NLP module is recommended to include a pre-processing block that filters non-essential text to better understand the text and increase the speed of the IQAS. The NLP module is recommended to include a part of speech processing block
