JP2004192491A

JP2004192491A - Question answering device, question answering program and recording medium recorded with same program

Info

Publication number: JP2004192491A
Application number: JP2002361678A
Authority: JP
Inventors: Tokuji Ikeno; 篤司池野
Original assignee: Oki Electric Industry Co Ltd
Current assignee: Oki Electric Industry Co Ltd
Priority date: 2002-12-13
Filing date: 2002-12-13
Publication date: 2004-07-08
Anticipated expiration: 2022-12-13
Also published as: JP4064802B2

Abstract

<P>PROBLEM TO BE SOLVED: To efficiently present an answer to a plurality of questions showing mutually related contents. <P>SOLUTION: This question answering device is provided with a question analysis part analyzing the relation of respective question contents of a plurality of question information to be fed in order, a document search part searching document information for searching the answer to the question of the question information from a document information group, an answer formation part extracting an answer element corresponding to an attribute of the answer required by the question from base document information and forming answer information using the answer element, an answer record part recording document identification information of the base document information linked with the answer information, and a record notification part notifying the document identification information to the document search part. When detecting that the question information is related to the question contents right before from the analysis result, the answer record part feeds the document identification information to the record notification part. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

【０００１】
【発明の属する技術分野】
本発明は、ユーザから自然言語で入力された質問事項に対し、所定の情報源から回答を検索し出力する情報検索機能を有する質問応答装置に関する。
【０００２】
【従来の技術】
コンピュータのユーザが情報検索のために入力した質問に対し、その回答を提示する手法として、後述する特許文献１の特開２００２−１３２８１１公報に開示された技術がある。この開示技術における質問応答システムは、例えば、「アメリカの大統領は誰ですか？」との質問文に対し、当該質問文の中の「アメリカ」および「大統領」を検索語として文書検索を行い、検索された文書から「クリントン」あるいは「プーチン」のような回答候補を得る。さらに、各回答候補について所定の手順でスコアリングを行った後、全ての回答候補の中から適正な回答を選定し、これを出力する。
【０００３】
【特許文献１】
特開２００２−１３２８１１公報
【０００４】
【発明が解決しようとする課題】
ところで、ユーザからの質問は、単発的なものばかりでなく、互いの内容に関連がある複数の質問となることがある。例えば、「ＡＢＣ社とＸＹＺ社とが合併したのはいつですか？」との質問に対する回答が提示された後、引き続いて、「その社長に就任したのは誰ですか？」のような質問を受けることがある。この質問に対する回答には、前者の質問における合併会社の社長の氏名が求められる。
【０００５】
しかしながら、このように直前の質問内容に関連がある質問文に対し、前記した開示技術では、自動的に「社長」および「就任」を検索語として検索することから、その検索結果には、前記合併会社に限らない各種企業の社長に関する文書が提示されることが予想される。その場合、ユーザは、提示された種々の文書の中から所望の文書を見つけ出さねばならない。従って、従来は、相互に関連した内容を有する複数の質問を行った場合、ユーザは適正な回答候補を円滑に取得することができないという問題があった。
【０００６】
本発明は、前記した課題に鑑みてなされたものであり、相互に関連する内容を示す複数の質問に対し、適正な回答を効率的に提示することができる質問応答装置を提供することを目的とする。
【０００７】
【課題を解決するための手段】
〈構成１〉
本発明に係る質問応答装置は、質問を示す質問情報を入力されたとき当該質問に対する回答を示す回答情報を出力する質問応答装置であって、順次的に供給される複数の質問情報のそれぞれについて各質問内容の関連を解析する質問解析部と、それぞれが文書識別情報により識別される文書情報を有する文書情報群から、質問情報の質問に対する回答を検索するための検索要素を含む文書情報を検索する文書検索部と、該文書検索部により検索された文書情報となる根拠文書情報から前記質問により要求される回答の属性に対応する回答要素を抽出し該抽出した回答要素を用いて回答情報を形成する回答形成部と、当該回答情報に前記根拠文書情報の文書識別情報を関連付けて記録する回答記録部と、該回答記録部の文書識別情報を前記文書検索部へ通知する記録通知部とを備え、前記回答記録部は、前記質問解析部から前記解析の結果を取得し、該取得した結果から前記質問情報と当該質問情報の直前の質問情報との間に内容の関連があることを検知したとき、前記直前の質問情報のための回答情報について記録した文書識別情報を前記記録通知部に供給する。
【０００８】
前記回答記録部は、前記質問解析部からの前記結果により前記質問情報と当該質問情報の直前の質問情報との間に内容の関連が無いことを検知したとき、前記記録通知部への前記文書識別情報の供給を停止することができる。
【０００９】
前記回答記録部は、前記記録通知部への前記文書識別情報の供給を停止すべく前記文書識別情報を破棄することができる。
【００１０】
また、前記回答形成部は、前記文書検索部からの根拠文書情報から前記回答要素が抽出できないとき、その旨を前記記録通知部に通知し、前記記録通知部が前記回答形成部から前記通知を受けたとき、当該根拠文書情報と異なる根拠文書情報の文書識別情報を前記回答記録部から取得し、該取得した文書識別情報を前記文書検索部へ供給することができる。
【００１１】
〈構成２〉
本発明に係る質問応答装置は、質問を示す質問情報を入力されたとき当該質問に対する回答を示す回答情報を出力する質問応答装置であって、順次的に供給される複数の質問情報のそれぞれについて各質問内容の関連を解析する質問解析部と、それぞれが文書識別情報により識別される文書情報を有する文書情報群から、前記質問情報の質問に対する回答を検索するための検索要素を含む文書情報を検索する文書検索部と、前記質問により要求される回答の属性に対応する回答要素を所定のスコアに基づき当該根拠文書情報から抽出し該抽出した回答要素を用いて回答情報を形成する回答形成部と、当該回答情報に前記根拠文書情報の文書識別情報を関連付けて記録する回答記録部と、該回答記録部に記録された文書識別情報毎に、前記スコアを調整するための重み値を設定する文書重み設定部とを備え、前記回答形成部は、前記文書検索部で検索された前記根拠文書情報の文書識別情報が、前記文書重み設定部で重み値を設定された文書識別情報に対応するとき、当該重み値に基づき回答要素を抽出する。
【００１２】
前記回答記録部は、前記質問解析部からの前記結果により前記質問情報と該質問情報の直前の質問情報との間に内容の関連が無いことを検知したとき、前記文書識別情報を破棄することができる。
【００１３】
また、前記回答記録部は、前記質問解析部からの前記結果により前記質問情報と該質問情報の直前の質問情報との間に内容の関連が無いことを検知したとき、前記文書識別情報を不活性化することができる。
【００１４】
前記重み設定部は、文書識別情報に対し、その記録順序に対応した所定の前記重み値を設定することができる。
【００１５】
〈構成３〉
本発明に係るプログラムは、コンピュータに質問を示す質問情報が供給されたとき当該質問に対する回答を示す回答情報を当該コンピュータに出力させる質問応答プログラムであって、複数の質問情報のそれぞれを順次に受け付け、受け付けた各質問情報について、その直前の質問情報との内容の関連を解析し、それぞれが文書識別情報により識別される文書情報を有する文書情報群から、前記質問に対する回答を検索するための検索要素を含む文書情報を検索し、前記検索された文書情報となる根拠文書情報から前記質問情報により要求される回答の属性に対応する回答要素を抽出し、該抽出した回答要素を用いて回答情報を形成し、当該回答情報に前記根拠文書情報の文書識別情報を関連付けて記録し、前記文書識別情報の記録後、前記質問情報と新たな質問情報との間に内容の関連があることを示す結果が得られたとき、前記記録された文書識別情報に対応する根拠文書情報から回答要素を抽出する。
【００１６】
〈構成４〉
本発明に係る記録媒体は、コンピュータに質問を示す質問情報が供給されたとき当該質問に対する回答を示す回答情報を当該コンピュータに出力させる質問応答プログラムを記録した記録媒体であって、前記質問応答プログラムは、複数の質問情報のそれぞれを順次に受け付け、受け付けた各質問情報について、その直前の質問情報との内容の関連を解析し、それぞれが文書識別情報により識別される文書情報を有する文書情報群から、前記質問に対する回答を検索するための検索要素を含む文書情報を検索し、前記検索された文書情報となる根拠文書情報から前記質問情報により要求される回答の属性に対応する回答要素を抽出し、該抽出した回答要素を用いて回答情報を形成し、当該回答情報に前記根拠文書情報の文書識別情報を関連付けて記録し、前記文書識別情報の記録後、前記質問情報と新たな質問情報との間に内容の関連があることを示す結果が得られたとき、前記記録された文書識別情報に対応する根拠文書情報から回答要素を抽出することを前記コンピュータに行わせる。
【００１７】
【発明の実施の形態】
以下、本発明の実施の形態を具体例を用いて説明する。
〈具体例１〉
図１は、本発明に係る質問応答装置の具体例１の構成を示すブロック図である。
具体例１の質問応答装置１０１は、ユーザに情報検索サービスを提供すべく、汎用のパーソナルコンピュータあるいは公共機関に設置された情報検索端末等に通信可能に接続され、これらの端末から質問情報を与えられたとき、その回答となる回答情報を出力する。
【００１８】
質問応答装置１０１は、図１に示すように、ユーザから入力された質問文を示す質問情報の解析により、検索のための検索要素、当該質問により要求される回答の属性となる回答タイプ、および、質問情報に関する後述する関連の有無を判定する質問解析部１０と、情報検索のための文書情報を保持する文書データベース１１と、該データベースから前記した検索要素が含まれる文書情報を検索する文書検索部１２と、該文書検索部で検索された文書情報である根拠文書情報から回答タイプに対応する回答要素を抽出し、その回答要素を用いて回答情報を形成する回答形成部１３と、当該回答情報を識別する回答ＩＤに前記根拠文書情報を識別する根拠文書ＩＤを関連付けて回答履歴として記録する回答記録部１４と、該回答記録部で記録された回答履歴を文書検索部１２へ供給する記録通知部１５とを備える。
【００１９】
質問解析部１０は、例えば、「ＡＢＣ社とＸＹＺ社が合併したのはいつですか？」の質問を示す質問情報ａ１を与えられたとき、従来の質問応答システムにおける手法と同様に、「ＡＢＣ社」、「ＸＹＺ社」および「合併」を示す検索要素を抽出すると共に、当該質問により求められる回答タイプが「年月日」であることを解析する。
【００２０】
また、本具体例の質問解析部１０は、与えられた質問の内容が直前の質問内容と関連するか否かを判定する。例えば、前記質問情報ａ１に引き続いて、当該質問に関連する「社長に就任したのは誰ですか？」との質問情報ａ２が与えられたとき、この質問情報ａ２の内容は質問情報ａ１の内容に関連するとの判定結果を出す。さらに、当該判定結果を回答記録部１４に通知する。
【００２１】
前記した内容の関連に関する判定にあたっては、例えば、順次入力される質問情報のそれぞれに、質問応答装置１０１の図示しない制御部により、直前の質問情報との関連を示す関連判定要素を埋め込む。質問解析部１０は、当該質問情報を受ける毎に、その関連判定要素を抽出する。具体的には、前記質問情報ａ１の次に入力された前記質問情報ａ２に、当該質問情報ａ１との間に内容の関連があることを示す関連判定要素を埋め込む。質問解析部１０は当該両質問の内容に関連があると判定する。また、前記質問情報ａ２の後に、「ＤＥＦ社の本社はどこにありますか？」との質問情報ｂ１が与えられたとき、当該質問情報ｂ１に、前記質問情報ａ２との関連が無いことを示す関連判定要素を埋め込む。質問解析部１０は、質問情報ｂ１の前記関連判定要素に基づき、質問情報ａ２および質問情報ｂ１が示す内容の関連が断たれたことを検知する。この方法は、質問内容の規則性が既知あるいは容易に推測できる場合に有効である。
【００２２】
また、関連性の判定に関する他の手法として、質問文の表現に関し所定のルールを設け、当該ルールに沿って質問情報の関連性を判定する手法を採用することができる。その場合、例えば、質問文の文頭に「その・・・は、」や、「それが・・・する」のように特定の単語を指示する表現が検出されたとき、あるいは、「先の・・・」や、「先程の（質問・回答）の・・・」のように過去の質問および回答自体またはその一部分を示すような表現が検出されたとき、質問に関連性があると判定することができる。他方、「ところで、・・・」や、「一方、・・・」のように話題の転換を示す表現が検出されたとき、あるいは、履歴をクリアするよう指示する内容や、質問ではないと判定される文が入力されたとき、質問の関連性が断たれたと判定するように設定することができる。また、質問を入力する間隔が所定時間を超えるとき、質問の関連が途切れたと判定することができる。
【００２３】
文書データベース１１は、それぞれが文書ＩＤで識別可能な文書情報を保持する。図１に示す例では、文書データベース１１が質問応答装置１０１内に設けられているが、この例に替えて、例えば質問応答装置１０１にネットワークを介して接続された文書データベース（１１）を適用することができる。
【００２４】
文書検索部１２は、質問解析部１０で得た検索要素を用いて、文書データベース１１から回答のための文書情報を検索するための検索クエリを作成し、当該クエリにより得られた文書情報となる根拠文書情報を、その文書ＩＤと共に回答形成部１３に供給する。
【００２５】
回答形成部１３は、従来の質問応答システムにおける回答抽出手順と同様に、質問情報の検索要素および回答タイプを用いて、検索スコアに基づき根拠文書情報から回答要素を抽出し、当該回答要素により回答情報を形成する。
また、本具体例の回答形成部１３は、回答情報の形成に利用した根拠文書情報の根拠文書ＩＤを当該回答情報の回答ＩＤに関連付け、これを回答履歴として回答記録部１４に通知する。
【００２６】
回答記録部１４は、回答形成部１３からの回答履歴を、図２に示すようなデータセットにまとめて保持する。
この回答記録部１４は、質問解析部１０に与えられた質問情報について、前記した関連性に関する判定結果を質問解析部１０から受け、その結果、当該質問情報がその直前の質問情報の内容と関連が無いことを検知した場合、既に蓄積した回答履歴を破棄あるいは不活性化する。また、前記結果が質問の関連を示すものである場合、当該質問の直前に受けた質問の回答ＩＤに対応する根拠文書ＩＤを記録通知部１５へ供給する。
【００２７】
記録通知部１５は、回答記録部１４からの根拠文書ＩＤを文書検索部１２または回答形成部１３からの要求に応じて供給する。
文書検索部１２からの前記要求は、処理中の質問情報が直前の質問に関連があるときに発生し、文書検索部１２が当該質問情報のために文書データベース１１を新たに検索することに代えて、当該質問情報の直前の質問情報に採用した根拠文書情報を利用すべく、その根拠文書ＩＤが要求される。文書検索部１２へ供給すべき根拠文書ＩＤは、前記したような直前の質問情報のそれに限らず、例えば一連の質問における最初の質問情報のために用いた根拠文書ＩＤを利用するなど、回答要素の抽出に有益と考えられるルールを適宜設定することができる。
【００２８】
また、回答形成部１３から記録通知部１５への根拠文書ＩＤの要求は、回答形成部１３が、直前の質問と関連がある質問に対する回答を作成するにあたり、文書検索部１２から供給された根拠文書情報で回答要素を抽出できないときに発生する。すなわち、回答形成部１３は、文書検索部１２から直前の質問情報の根拠文書情報を取得しても、当該情報に適正な回答要素が存在しない場合、他の根拠文書情報を取得するために、回答履歴を遡って根拠文書情報を供給するよう回答形成部１３に要求する。この要求により採用される根拠文書ＩＤは、例えば、直前の質問情報から履歴を１つ遡った質問情報の根拠文書ＩＤ、あるいは、その履歴における最初の根拠文書ＩＤ等、適宜設定することができる。
【００２９】
回答形成部１３が回答情報を形成する手順を図３に示すフローチャートに沿って説明する。
回答形成部１３は、前記したように、質問情報の検索要素および回答タイプを用いて、従来の質問応答システムにおける回答抽出手順と同様に、検索スコアに基づき根拠文書情報から抽出すべき回答要素を探索する(ステップＳ１)。
【００３０】
さらに、前記根拠文書情報からの回答要素の抽出に成功したか否か判定し(ステップＳ２)、成功した場合、抽出した回答要素を用いて回答情報を形成すると共に、当該根拠文書情報の根拠文書ＩＤを当該回答情報の回答ＩＤに関連付け、これを回答履歴として回答記録部１４に通知する(ステップＳ３)。形成された回答情報はユーザに提示される(ステップＳ４)。
【００３１】
他方、前記根拠文書情報から回答要素を抽出できなかった場合、回答形成部１３は、その旨を記録通知部１５に通知し(ステップＳ５)、回答履歴に他の根拠文書ＩＤの記録があるか否かの判定を促す(ステップＳ６)。参照すべき他の根拠文書ＩＤがあれば、その根拠文書情報を取得すべく、当該ＩＤを文書検索部１２へ通知するよう記録通知部１５に要求し(ステップＳ７)、一旦、処理を終了する。その後、文書検索部１２から回答形成部１３へ新たな根拠文書情報が供給される。
【００３２】
また、参照すべき他の根拠文書ＩＤが回答履歴に存在しなければ、当該質問に対しユーザに供給すべき情報が無い旨の回答を用意する(ステップＳ８)。その場合、例えば「ご要望の情報が見つかりません。」のような案内がユーザに通知される(ステップＳ４)。尚、このように回答履歴に他の根拠文書ＩＤが無い場合、その時点で検索処理を終了することに代えて、回答履歴に関連しない新たな文書情報を文書データベース１１から取得し、その文書情報から回答要素を探索するようにしてもよい。
【００３３】
具体例１の質問応答装置１０１によれば、入力された質問情報の内容と、その直前の質問内容との間に関連があると判定したとき、新たに文書検索を行うことに代えて、既に利用した根拠文書情報を優先的に参照することから、一連の質問に対する回答の処理時間を短縮することができる。しかも、関連する質問の回答に用いられた前記根拠文書情報は、処理中の質問情報にとって有益な回答要素を含む可能性が高いことから、ユーザに対し適正な回答情報を提示することができる。
【００３４】
〈具体例２〉
図４は、本発明に係る質問応答装置の具体例２の構成を示すブロック図である。
具体例２の質問応答装置１０２は、図４に示すように、前記した具体例１の質問応答装置１０１における質問解析部１０、文書データベース１１、文書検索部１２、回答形成部１３および回答記録部１４と基本的な機能を同じくする、質問解析部２０、文書データベース２１、文書検索部２２、回答形成部２３および回答記録部２４を備える。
【００３５】
質問応答装置１０２には、前記構成に加えて、回答形成部２３による回答要素の抽出に際し用いられる検索スコアを調整するための重み値を設定する文書重み設定部２６が設けられている。文書重み設定部２６は、回答記録部２４の回答履歴に記録されている根拠文書ＩＤを全て取得し、取得した各ＩＤに所定の重み値を対応付ける。図５は、文書重み設定部２６で保持される重み設定リストの一例を示す。
【００３６】
文書重み設定部２６が行う重み値の設定方法としては、例えば、入力された質問情報の直前の質問情報に対応する根拠文書ＩＤ、すなわち回答履歴における最新の根拠文書ＩＤに対して、最大の重み値を付与し、履歴を遡る毎に重み値を所定の割合で漸減させるように設定することができる。また、この方法に代えて、回答履歴において最初に記録された根拠文書ＩＤに最大の重み値を設定し、当該重み値より低い値を他のＩＤに均一に付与することもできる。
【００３７】
なお、本具体例では、重み値が各根拠文書ＩＤに１つずつ設定されるが、例えば図２の回答履歴における「3291.txt」のように、履歴中に複数回現れる根拠文書ＩＤがある場合、当該ＩＤに対する重み値としては、設定可能な最大値を付与する、あるいは、加算または乗算により値を累積させる等、適宜調整し、当該根拠文書ＩＤに単一の重み値が設定されるようにする。
【００３８】
具体例２の質問応答装置１０２の動作を説明する。ここでは、質問応答装置１０２に互いに内容の関連を有する質問情報が入力されるものとする。
質問解析部２０は、具体例１の質問解析部１０と同様な手順で当該質問情報を解析し、その検索要素および回答タイプを文書検索部２２へ供給すると共に、質問の関連性の判定結果を回答記録部２４へ通知する。
【００３９】
文書検索部２２は、前記検索要素を用いて検索クエリを作成し、当該クエリにより得られた全ての根拠文書情報を、それらの文書ＩＤと共に回答形成部２３に供給する。本具体例の文書検索部２２は、質問を受ける都度、その質問内容が直前の質問内容に関連するか否かに関わらず、文書データベース２１から根拠文書情報を検索する。
【００４０】
回答形成部２３は、文書検索部２２から取得した根拠文書情報の根拠文書ＩＤに関し、その重み値を問い合わせる。一例として、前記根拠文書情報が「3291.txt」を根拠文書ＩＤとする場合、文書重み設定部２６は、図５に示す前記した重み設定リストを参照し、当該ＩＤの重み値である「1.5」を回答形成部２３へ通知する。
【００４１】
回答形成部２３は、前記重み値を用いて、対応する根拠文書情報の所定の検索スコアを調整し、当該スコアに基づき回答要素を抽出する。そして、具体例１の回答形成部１３の動作手順と同様に、前記回答要素から回答情報を形成して出力すると共に、回答記録部２４へ回答履歴を通知する。
【００４２】
具体例２の質問応答装置１０２によれば、回答履歴に記録された各根拠文書ＩＤに重み値を設定し、対応する根拠文書情報の検索スコアに当該重み値を反映させることから、質問内容が直前のものに関連する質問を処理するうえで有益な文書情報を優先的に採用することができる。これにより、回答要素を抽出する精度を高めることができ、ユーザに対し回答情報を効率的に提示することが可能となる。
【００４３】
前記した各具体例では、本発明に係る質問応答装置を情報検索装置に適用した例を示したが、この他に、例えば質問回答形式に限らない対話を行うための対話装置に適用することができる。
【００４４】
また、前記した各具体例の動作手順に沿った質問応答プログラムをコンピュータに導入することにより、種々のコンピュータを質問応答装置として動作させることができる。さらに、このプログラムを記録媒体に記録し、当該記録媒体をコンピュータに利用することができる。
【００４５】
【発明の効果】
本発明に係る質問応答装置によれば、順次的に供給される質問情報の質問内容に関連があると判定したとき、既に利用した根拠文書情報を優先的に参照することから、回答の処理時間を短縮することができる。また、既に利用した各根拠文書情報に対し予め重み値を設定し、処理中の検索スコアに当該重み値を反映させることから、直前の質問内容と関連がある質問にとって有益な文書情報を優先的に採用することができる。従って、ユーザに対し回答情報を効率的に提示することが可能となる。
【図面の簡単な説明】
【図１】本発明に係る質問応答装置の具体例１の構成を示すブロック図である。
【図２】回答記録部の回答履歴を説明するための説明図である。
【図３】具体例１の回答形成部の動作を示すフローチャートである。
【図４】本発明に係る質問応答装置の具体例２の構成を示すブロック図である。
【図５】文書重み設定部の重み設定リストを説明するための説明図である。
【符号の説明】
１０１、１０２質問応答装置
１０、２０質問解析部
１１、２１文書データベース
１２、２２文書検索部
１３、２３回答形成部
１４、２４回答記録部
１５記録通知部
２６文書重み設定部[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a question answering apparatus having an information search function for searching and outputting an answer from a predetermined information source for a question item input in a natural language from a user.
[0002]
[Prior art]
As a method of presenting an answer to a question input by a computer user for information search, there is a technique disclosed in Japanese Patent Application Laid-Open No. 2002-132811 of Patent Document 1 described later. The question answering system in the disclosed technology performs, for example, a document search for a question sentence "Who is the United States of America?", Using "America" and "President" in the question sentence as search terms, Answer candidates such as "Clinton" or "Putin" are obtained from the retrieved documents. Further, after scoring each answer candidate according to a predetermined procedure, an appropriate answer is selected from all the answer candidates and is output.
[0003]
[Patent Document 1]
JP-A-2002-132811
[Problems to be solved by the invention]
By the way, the questions from the user are not only one-shot questions but may be a plurality of questions related to each other. For example, after answering the question "When was ABC and XYZ merged?", A question like "Who was the president?" May be received. The answer to this question requires the name of the president of the merged company in the first question.
[0005]
However, for the question sentence related to the content of the immediately preceding question, the disclosed technology automatically searches for “President” and “Inauguration” as search terms. It is expected that documents related to the presidents of various companies not limited to the merged company will be presented. In that case, the user must find a desired document from the various documents presented. Therefore, conventionally, there has been a problem that when a plurality of questions having mutually related contents are performed, the user cannot smoothly obtain appropriate answer candidates.
[0006]
The present invention has been made in view of the above-described problems, and has as its object to provide a question answering apparatus capable of efficiently presenting appropriate answers to a plurality of questions indicating mutually related contents. And
[0007]
[Means for Solving the Problems]
<Configuration 1>
A question answering apparatus according to the present invention is a question answering apparatus that outputs answer information indicating an answer to a question when question information indicating a question is input, and for each of a plurality of question information sequentially supplied. A question analysis unit that analyzes the relationship between the contents of each question, and a document information group that includes document information identified by the document identification information, and searches for document information including a search element for searching for an answer to the question of the question information. A document search unit that extracts the answer element corresponding to the attribute of the answer requested by the question from the basis document information that is the document information searched by the document search unit, and uses the extracted answer element to convert the answer information. An answer forming unit for forming, an answer recording unit for recording the answer information in association with the document identification information of the basis document information, and a document identification information of the answer recording unit for the document. A record notifying unit for notifying the search unit, the answer recording unit obtains the result of the analysis from the question analysis unit, and obtains the question information and the question information immediately before the question information from the obtained result. When it is detected that the contents are related to each other, the document identification information recorded on the answer information for the immediately preceding question information is supplied to the recording notifying unit.
[0008]
The answer recording unit, when detecting that there is no content relationship between the question information and the question information immediately before the question information based on the result from the question analysis unit, the document to the record notification unit The supply of the identification information can be stopped.
[0009]
The answer recording unit may discard the document identification information so as to stop supplying the document identification information to the record notification unit.
[0010]
Further, when the answer element cannot be extracted from the basis document information from the document search unit, the answer forming unit notifies the recording notifying unit of the fact, and the recording notifying unit transmits the notification from the answer forming unit. When receiving the document identification information, the document identification information of the basis document information different from the basis document information can be acquired from the answer recording unit, and the acquired document identification information can be supplied to the document search unit.
[0011]
<Configuration 2>
A question answering apparatus according to the present invention is a question answering apparatus that outputs answer information indicating an answer to a question when question information indicating a question is input, and for each of a plurality of question information sequentially supplied. A question analysis unit that analyzes the relationship between the contents of each question, and a document information group that includes a search element for searching for an answer to the question of the question information from a document information group having document information each identified by the document identification information. A document search unit to be searched, and an answer forming unit that extracts answer elements corresponding to the attribute of the answer requested by the question from the base document information based on a predetermined score and forms answer information using the extracted answer elements A response recording unit that records the response information in association with the document identification information of the base document information; and, for each document identification information recorded in the response recording unit, A document weight setting unit for setting a weight value for adjusting the document identification information of the basis document information retrieved by the document retrieval unit. , The response element is extracted based on the weight value.
[0012]
The answer recording unit discards the document identification information when detecting that there is no content relationship between the question information and the question information immediately before the question information based on the result from the question analysis unit. Can be.
[0013]
Further, when the answer recording unit detects that there is no content relationship between the question information and the question information immediately before the question information based on the result from the question analysis unit, the answer recording unit does not include the document identification information. Can be activated.
[0014]
The weight setting unit can set the predetermined weight value corresponding to the recording order in the document identification information.
[0015]
<Configuration 3>
A program according to the present invention is a question answering program that, when question information indicating a question is supplied to a computer, causes the computer to output answer information indicating an answer to the question, and sequentially receives each of the plurality of pieces of question information. For each of the received question information, analyzing the relationship between the content of the received question information and the immediately preceding question information, and searching for an answer to the question from a document information group having document information each identified by the document identification information. Searching for document information including an element, extracting an answer element corresponding to the attribute of the answer requested by the question information from the basis document information that is the searched document information, and using the extracted answer element to set the answer information. Is formed, and the response information is recorded in association with the document identification information of the base document information, and after the recording of the document identification information, the If the result indicating that there is associated content of between question information and the new question information is obtained, extracting the reply element from basis document information corresponding to the recorded document identification information.
[0016]
<Configuration 4>
The recording medium according to the present invention is a recording medium recording a question answering program for causing a computer to output answer information indicating an answer to the question when question information indicating the question is supplied to the computer, wherein the question answering program Is a document information group having document information each of which sequentially receives a plurality of pieces of question information, analyzes the relationship between the content of the received question information and the content of the immediately preceding question information, and each of which has document information identified by the document identification information. Searching for document information including a search element for searching for an answer to the question, and extracting an answer element corresponding to the attribute of the answer requested by the question information from the basis document information that is the searched document information Then, answer information is formed using the extracted answer elements, and the document identification information of the base document information is associated with the answer information. After recording the document identification information, when a result indicating that there is a content relationship between the question information and the new question information is obtained, a basis corresponding to the recorded document identification information is obtained. The computer is caused to extract an answer element from document information.
[0017]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described using specific examples.
<Specific example 1>
FIG. 1 is a block diagram showing the configuration of Example 1 of the question answering apparatus according to the present invention.
The question answering apparatus 101 of the specific example 1 is communicably connected to a general-purpose personal computer or an information search terminal installed in a public institution to provide an information search service to a user, and gives question information from these terminals. When it is received, it outputs answer information as the answer.
[0018]
As shown in FIG. 1, the question answering apparatus 101 analyzes the question information indicating the question sentence input by the user, and searches for a search element for search, an answer type serving as an attribute of an answer required by the question, and A question analysis unit 10 for determining whether or not the question information is related as described below, a document database 11 holding document information for information search, and a document search for searching the database for document information including the search element described above. An answer forming unit 13 for extracting an answer element corresponding to the answer type from the basis document information which is the document information searched by the document searching unit, and forming the answer information using the answer element; An answer recording unit 14 for associating the answer document ID for identifying the information with the basis document ID for identifying the basis document information and recording the answer history; The answer history and a recording notification unit 15 supplies the document retrieval unit 12.
[0019]
For example, when given the question information a1 indicating the question "When was ABC and XYZ merged?", The question analysis unit 10 sets "ABC" in the same manner as in the conventional question answering system. A search element indicating “company”, “XYZ company”, and “merger” is extracted, and it is analyzed that the answer type required by the question is “date”.
[0020]
In addition, the question analysis unit 10 of this specific example determines whether or not the content of the given question is related to the content of the immediately preceding question. For example, following the question information a1, when question information a2 related to the question, "Who became the president?" Is given, the content of the question information a2 is the content of the question information a1. The result of the determination that is related to. Further, the determination result is notified to the answer recording unit 14.
[0021]
In the above-described determination regarding the relationship between the contents, for example, a relationship determination element indicating the relationship with the immediately preceding question information is embedded in each of the sequentially input question information by the control unit (not illustrated) of the question answering apparatus 101. Each time the question analyzer 10 receives the question information, it extracts the relevant determination element. Specifically, a relation determination element indicating that the contents are related to the question information a1 is embedded in the question information a2 input after the question information a1. The question analysis unit 10 determines that the contents of the two questions are related. Also, when question information b1 "where is the head office of DEF company?" Is given after question information a2, a relation indicating that question information b1 is not related to question information a2. Embed the judgment element. The question analysis unit 10 detects that the association between the content indicated by the question information a2 and the content indicated by the question information b1 has been broken based on the association determination element of the question information b1. This method is effective when the regularity of the question content is known or can be easily estimated.
[0022]
Further, as another method for determining the relevance, a method may be adopted in which a predetermined rule is set for the expression of the question text and the relevance of the question information is determined according to the rule. In that case, for example, when an expression indicating a specific word such as "that ..." or "it ..." is detected at the beginning of the question, If an expression indicating a past question and the answer itself or a part thereof, such as ".." or "(Question / Answer) ...", is detected, it is determined that the question is relevant. be able to. On the other hand, when an expression indicating a change in topic, such as “By the way,” or “One side,...”, Is detected, or it is determined that the content is an instruction to clear the history or is not a question. When a sentence is input, it can be set to determine that the relevance of the question has been cut. When the interval of inputting the questions exceeds a predetermined time, it can be determined that the association of the questions has been interrupted.
[0023]
The document database 11 holds document information that can be identified by a document ID. In the example shown in FIG. 1, the document database 11 is provided in the question answering apparatus 101. Instead of this example, for example, a document database (11) connected to the question answering apparatus 101 via a network is applied. be able to.
[0024]
The document search unit 12 uses the search elements obtained by the question analysis unit 10 to create a search query for searching document information for an answer from the document database 11, and becomes the document information obtained by the query. The basis document information is supplied to the answer forming unit 13 together with the document ID.
[0025]
The answer forming unit 13 extracts the answer element from the basis document information based on the search score using the search element and the answer type of the question information, as in the answer extraction procedure in the conventional question answering system, and answers with the answer element. Form information.
In addition, the answer forming unit 13 of this specific example associates the basis document ID of the basis document information used for forming the answer information with the answer ID of the answer information, and notifies the answer recording unit 14 of this as the answer history.
[0026]
The answer recording unit 14 collectively holds the answer histories from the answer forming unit 13 in a data set as shown in FIG.
The answer recording unit 14 receives, from the question analysis unit 10, the determination result regarding the relevance described above for the question information given to the question analysis unit 10, and as a result, the question information is related to the content of the immediately preceding question information. When it is detected that there is no response history, the already accumulated answer history is discarded or inactivated. When the result indicates the relation of the question, the base document ID corresponding to the answer ID of the question received immediately before the question is supplied to the record notification unit 15.
[0027]
The record notifying unit 15 supplies the basis document ID from the answer recording unit 14 in response to a request from the document search unit 12 or the answer forming unit 13.
The request from the document search unit 12 occurs when the question information being processed is related to the immediately preceding question, and instead of the document search unit 12 newly searching the document database 11 for the question information, Then, the basis document ID is required to use the basis document information adopted as the question information immediately before the question information. The base document ID to be supplied to the document search unit 12 is not limited to that of the immediately preceding question information, but may be an answer element such as a base document ID used for the first question information in a series of questions. Can be appropriately set as rules that are considered to be useful for the extraction of.
[0028]
Further, the request of the basis document ID from the answer forming unit 13 to the record notifying unit 15 is based on the basis supplied from the document searching unit 12 when the answer forming unit 13 creates an answer to the question related to the immediately preceding question. Occurs when answer elements cannot be extracted from document information. That is, even if the answer forming unit 13 obtains the basis document information of the immediately preceding question information from the document search unit 12, if there is no appropriate answer element in the information, the answer forming unit 13 obtains another basis document information. It requests the answer forming unit 13 to supply the base document information by going back the answer history. The basis document ID adopted by this request can be set as appropriate, for example, the basis document ID of the question information that is traced back from the immediately preceding question information by one, or the first basis document ID in the history.
[0029]
The procedure by which the answer forming unit 13 forms answer information will be described with reference to the flowchart shown in FIG.
As described above, the answer forming unit 13 uses the search element and the answer type of the question information to determine the answer element to be extracted from the basis document information based on the search score, as in the answer extraction procedure in the conventional question answering system. Search (step S1).
[0030]
Further, it is determined whether or not the extraction of the answer element from the basis document information is successful (step S2). If the answer element is successful, the answer information is formed using the extracted answer element and the basis document of the basis document information is determined. The ID is associated with the answer ID of the answer information, and this is notified to the answer recording unit 14 as an answer history (step S3). The formed answer information is presented to the user (step S4).
[0031]
On the other hand, when the answer element cannot be extracted from the basis document information, the answer forming unit 13 notifies the record notifying unit 15 of the fact (step S5), and determines whether there is a record of another basis document ID in the answer history. The user is prompted to determine whether or not there is (Step S6). If there is another basis document ID to be referred to, the recording request unit 15 is requested to notify the document search unit 12 of the ID in order to acquire the basis document information (step S7), and the process is temporarily terminated. . After that, new document information is supplied from the document search unit 12 to the answer forming unit 13.
[0032]
If another base document ID to be referred to does not exist in the answer history, a reply is prepared that there is no information to be supplied to the user for the question (step S8). In this case, the user is notified of a guidance such as "The requested information cannot be found" (step S4). If there is no other base document ID in the answer history as described above, instead of terminating the search process at that point, new document information not related to the answer history is acquired from the document database 11 and the document information is acquired. May be searched for an answer element.
[0033]
According to the question answering apparatus 101 of the specific example 1, when it is determined that there is a relationship between the content of the input question information and the content of the immediately preceding question, instead of performing a new document search, Since the basis document information used is preferentially referred to, it is possible to reduce the processing time of the answer to a series of questions. Moreover, since the basis document information used for answering the related question has a high possibility of including an answer element useful for the question information being processed, it is possible to present appropriate answer information to the user.
[0034]
<Specific example 2>
FIG. 4 is a block diagram showing the configuration of Example 2 of the question answering device according to the present invention.
As shown in FIG. 4, the question answering apparatus 102 of the specific example 2 includes a question analyzing unit 10, a document database 11, a document searching unit 12, an answer forming unit 13, and an answer recording unit in the question answering apparatus 101 of the specific example 1 described above. It has a question analysis unit 20, a document database 21, a document search unit 22, an answer formation unit 23, and an answer recording unit 24 having the same basic functions as those of the.
[0035]
In addition to the above configuration, the question answering apparatus 102 is provided with a document weight setting unit 26 that sets a weight value for adjusting a search score used when the answer forming unit 23 extracts an answer element. The document weight setting unit 26 acquires all the basis document IDs recorded in the answer history of the answer recording unit 24, and associates each acquired ID with a predetermined weight value. FIG. 5 shows an example of the weight setting list held by the document weight setting unit 26.
[0036]
As a method of setting the weight value performed by the document weight setting unit 26, for example, the maximum weight is assigned to the basis document ID corresponding to the question information immediately before the input question information, that is, the latest basis document ID in the answer history. A value can be given and the weight value can be set so as to gradually decrease at a predetermined rate every time the history is traced. Instead of this method, it is also possible to set a maximum weight value for the base document ID recorded first in the answer history, and to uniformly assign a value lower than the weight value to other IDs.
[0037]
In this specific example, one weight value is set for each basis document ID, but there is a basis document ID that appears multiple times in the history, such as “3291.txt” in the answer history in FIG. In this case, as a weight value for the ID, a maximum value that can be set is given, or a value is accumulated as appropriate by adding or multiplying, so that a single weight value is set for the base document ID. To
[0038]
The operation of the question answering apparatus 102 of the specific example 2 will be described. Here, it is assumed that question information having a content relationship with each other is input to the question answering device 102.
The question analysis unit 20 analyzes the question information in the same procedure as the question analysis unit 10 of the specific example 1, supplies the search element and the answer type to the document search unit 22, and outputs the result of determining the relevance of the question. Notify the answer recording unit 24.
[0039]
The document search unit 22 creates a search query using the search elements, and supplies all the base document information obtained by the query to the answer forming unit 23 together with their document IDs. Each time the question is received, the document search unit 22 of this specific example searches the document database 21 for the base document information regardless of whether the content of the question is related to the content of the immediately preceding question.
[0040]
The answer forming unit 23 inquires about the weight value of the basis document ID of the basis document information acquired from the document search unit 22. As an example, when the basis document information is “3291.txt” as the basis document ID, the document weight setting unit 26 refers to the above-described weight setting list shown in FIG. To the answer forming unit 23.
[0041]
The answer forming unit 23 adjusts a predetermined search score of the corresponding basis document information using the weight value, and extracts an answer element based on the score. Then, in the same manner as the operation procedure of the answer forming unit 13 of the specific example 1, the answer information is formed and output from the answer elements, and the answer record unit 24 is notified of the answer history.
[0042]
According to the question answering apparatus 102 of the specific example 2, a weight value is set for each basis document ID recorded in the answer history, and the weight value is reflected in the search score of the corresponding basis document information. Document information useful in processing a question related to the immediately preceding one can be preferentially adopted. Thereby, the accuracy of extracting the answer element can be improved, and the answer information can be efficiently presented to the user.
[0043]
In each of the specific examples described above, the example in which the question answering device according to the present invention is applied to the information search device is shown. it can.
[0044]
In addition, by introducing a question answering program in accordance with the operation procedure of each of the above specific examples into a computer, various computers can be operated as a question answering device. Further, this program can be recorded on a recording medium, and the recording medium can be used for a computer.
[0045]
【The invention's effect】
According to the question answering apparatus of the present invention, when it is determined that the question contents of the question information sequentially supplied are related, the reference document information that has already been used is preferentially referred to. Can be shortened. In addition, since a weight value is set in advance for each basis document information that has already been used and the weight value is reflected in the search score being processed, document information that is useful for a question related to the content of the immediately preceding question is given priority. Can be adopted. Therefore, it is possible to efficiently present answer information to the user.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a configuration of a specific example 1 of a question answering apparatus according to the present invention.
FIG. 2 is an explanatory diagram for explaining an answer history of an answer recording unit.
FIG. 3 is a flowchart illustrating an operation of an answer forming unit according to the first embodiment.
FIG. 4 is a block diagram showing a configuration of a specific example 2 of the question answering device according to the present invention.
FIG. 5 is an explanatory diagram for explaining a weight setting list of a document weight setting unit.
[Explanation of symbols]
101, 102 Question answering device 10, 20 Question analysis unit 11, 21 Document database 12, 22 Document search unit 13, 23 Answer formation unit 14, 24 Answer recording unit 15 Record notifying unit 26 Document weight setting unit

Claims

A question answering device that outputs answer information indicating an answer to the question when question information indicating the question is input,
A question analysis unit that analyzes the relation of each question content for each of a plurality of question information sequentially supplied, and answers to questions of the question information from a document information group having document information each identified by document identification information A document search unit for searching for document information including a search element for searching for, and an answer element corresponding to an attribute of an answer requested by the question from basis document information serving as document information searched by the document search unit. An answer forming unit that extracts and forms answer information using the extracted answer element; an answer recording unit that records the answer information in association with the document identification information of the base document information; and a document identification information of the answer recording unit. And a record notifying unit for notifying the document search unit, the answer recording unit acquires the result of the analysis from the question analysis unit, and obtains the question information from the acquired result. When detecting that there is a relationship between the contents of the question information and the immediately preceding question information, the document identification information recorded with respect to the answer information for the immediately preceding question information is supplied to the recording notification unit. Question answering device.

The answer recording unit, when detecting that there is no relationship between the content of the question information and the question information immediately before the question information based on the result from the question analysis unit, the document to the record notification unit 2. The question answering device according to claim 1, wherein the supply of the identification information is stopped.

3. The question answering device according to claim 2, wherein the answer recording unit discards the document identification information so as to stop supplying the document identification information to the record notification unit.

The answer forming unit, when the answer element cannot be extracted from the basis document information from the document search unit, notifies the record notifying unit to that effect,
The record notifying unit, when receiving the notification from the answer forming unit, acquires the document identification information of the basis document information different from the basis document information from the answer recording unit, and searches the acquired document identification information for the document. The question answering device according to claim 1, which is supplied to a section.

A question answering device that outputs answer information indicating an answer to the question when question information indicating the question is input,
A question analysis unit that analyzes the association of each question content with respect to each of a plurality of question information sequentially supplied, and a document information group having document information each identified by document identification information. A document search unit that searches for document information including a search element for searching for an answer, and an answer element corresponding to an attribute of the answer requested by the question is extracted and extracted from the base document information based on a predetermined score. An answer forming unit that forms answer information using answer elements, an answer recording unit that records the answer information in association with the document identification information of the basis document information, and a document identification information recorded in the answer recording unit. A document weight setting unit for setting a weight value for adjusting the score,
The answer forming unit, when the document identification information of the basis document information searched by the document search unit corresponds to the document identification information for which a weight value has been set by the document weight setting unit, provides an answer based on the weight value. A question answering device characterized by extracting an element.

The answer recording unit discards the document identification information when detecting that there is no content relationship between the question information and the question information immediately before the question information based on the result from the question analysis unit. Item 5. The question answering device according to Item 5.

The answer recording unit deactivates the document identification information when detecting that there is no content relationship between the question information and the question information immediately before the question information based on the result from the question analysis unit. 6. The question answering device according to claim 5, wherein

6. The question answering device according to claim 5, wherein the document weight setting unit sets the predetermined weight value corresponding to the recording order in the document identification information.

A question answering program that causes the computer to output answer information indicating an answer to the question when question information indicating the question is supplied to the computer,
Each of the plurality of pieces of question information is sequentially received, and for each of the received pieces of question information, the relationship between the content and the immediately preceding question information is analyzed,
From each document information group having document information identified by the document identification information, searching for document information including a search element for searching for an answer to the question,
Extracting an answer element corresponding to the attribute of the answer requested by the question information from the basis document information that is the searched document information, forming answer information using the extracted answer element,
The answer information is recorded in association with the document identification information of the basis document information,
After recording the document identification information, when a result indicating that there is a content relationship between the question information and the new question information is obtained, a response is made from the basis document information corresponding to the recorded document identification information. A question answering program characterized by extracting elements.

A recording medium recording a question answering program that causes the computer to output answer information indicating an answer to the question when question information indicating the question is supplied to the computer,
The question answering program,
Each of the plurality of pieces of question information is sequentially received, and for each of the received pieces of question information, the relationship between the content and the immediately preceding question information is analyzed,
From each document information group having document information identified by the document identification information, searching for document information including a search element for searching for an answer to the question,
Extracting an answer element corresponding to the attribute of the answer requested by the question information from the basis document information that is the searched document information, forming answer information using the extracted answer element,
The answer information is recorded in association with the document identification information of the basis document information,
After recording the document identification information, when a result indicating that there is a content relationship between the question information and the new question information is obtained, a response is made from the basis document information corresponding to the recorded document identification information. A recording medium characterized by causing the computer to extract an element.