Asking Questions: an Innovative Way to Interact with Oral History Archives

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F49777513%3A23520%2F23%3A43969939" target="_blank" >RIV/49777513:23520/23:43969939 - isvavai.cz</a>

Výsledek na webu

<a href="http://svk.fav.zcu.cz/download/proceedings_svk_2023.pdf" target="_blank" >http://svk.fav.zcu.cz/download/proceedings_svk_2023.pdf</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Asking Questions: an Innovative Way to Interact with Oral History Archives

Popis výsledku v původním jazyce

Oral history archives are a large source of historical knowledge. Different institutions are collecting interviews and testimonies related to major historical topics. Listening to huge quantities of audio materials is impractical for a common user. On the other side, the testimonies provided by the interviewees were intended to give evidence of historical events, and it is not ethical to change the meaning or cherry-pick single facts from a given interview. Many efforts to provide access to such archives were proposed in recent research works, including speech-to-text technologies and spoken-term detection methods. Other approaches used traditional information retrieval methods. However, the most used method for accessing the archives is by using a keyword search in automatically / manually created transcripts.This paper proposes an innovative approach that integrates the above-mentioned technologies. The approach generates a new question-answer structure on top of the existing interview transcripts. The questions are automatically generated together with related answers for a specific interview fragment. The questions can be indexed and time-aligned with the audio so the user can quickly get to interesting parts of the interview. The questions complement the interviewer and are useful in passages where only the interviewee speaks. It is important to stress that the questions do not change the meaning of the testimony because the related parts of the original interview are presented as an answer. In contrast to the question-answering methods, we use the term asking questions for our approach (abbreviated as AQ in contrast to question answering – QA).

Název v anglickém jazyce

Asking Questions: an Innovative Way to Interact with Oral History Archives

Popis výsledku anglicky

Oral history archives are a large source of historical knowledge. Different institutions are collecting interviews and testimonies related to major historical topics. Listening to huge quantities of audio materials is impractical for a common user. On the other side, the testimonies provided by the interviewees were intended to give evidence of historical events, and it is not ethical to change the meaning or cherry-pick single facts from a given interview. Many efforts to provide access to such archives were proposed in recent research works, including speech-to-text technologies and spoken-term detection methods. Other approaches used traditional information retrieval methods. However, the most used method for accessing the archives is by using a keyword search in automatically / manually created transcripts.This paper proposes an innovative approach that integrates the above-mentioned technologies. The approach generates a new question-answer structure on top of the existing interview transcripts. The questions are automatically generated together with related answers for a specific interview fragment. The questions can be indexed and time-aligned with the audio so the user can quickly get to interesting parts of the interview. The questions complement the interviewer and are useful in passages where only the interviewee speaks. It is important to stress that the questions do not change the meaning of the testimony because the related parts of the original interview are presented as an answer. In contrast to the question-answering methods, we use the term asking questions for our approach (abbreviated as AQ in contrast to question answering – QA).

Druh

O - Ostatní výsledky

CEP obor

—

OECD FORD obor

20205 - Automation and control systems

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2023

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů