「英語で読ませた時でさえ、マンガの要約に集合知は有効か？」
研究者らの過去の実験では、１００コマ程度のストーリーマンガを大学生に与え、５％～２０％のコマに要約させ、選択率の高い上位ｎ％のコマを抜き出すと、ｎをいくつにしても良い要約になった。つまりマンガの要約に、集合知は非常に良く機能した。しかし集合知はどれほど頑健だろうか？ たとえば、同じマンガを英語に翻訳して読ませた時でさえ、なお集合知は有効であり続けるのだろうか？ 60名の大学生に実験を行った結果を報告する。
"Is collective intelligence effective in summarizing Manga even when it is written in English"
According to a past experiment by researchers, when a story manga with about 100 cutscenes was given to university students and let them summarize it into 5% to 20% of cutscenes and then pick out n% of those cutscenes in the order of selected frequency, it became nice summary regardless of what number n is. In other words, collective intelligence worked extremely well in summarizing manga. But how reliable collective intelligence is? Will collective intelligence continue to be effective even if the same manga translated into English is asked to be read? I will report the results of the experiment having 60 university students participated in the study.
