Estudo constata que homens são representados quatro vezes mais que mulheres na literatura
Pesquisadores da Escola de Engenharia USC Viterbi utilizaram tecnologias de IA para concluir que personagens masculinos são quatro vezes mais prevalentes na literatura do que personagens femininos.
![](data:image/jpeg;base64,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)
Pixabay
Pesquisadores da Escola de Engenharia USC Viterbi utilizaram tecnologias de IA para concluir que personagens masculinos são quatro vezes mais prevalentes na literatura do que personagens femininos.
Mayank Kejriwal, lider de pesquisa do Instituto de Ciências da Informação da USC (ISI), foi inspirado pelo trabalho atual sobre preconceitos implacitos de gaªnero e sua própria experiência em processamento de linguagem natural (PNL). Enquanto muitos estudos publicados pesquisam e analisam os aspectos qualitativos da representação feminina na literatura e na madia, a pesquisa de Kejriwal fez uso particularmente de seus pontos fortes oscoleta de dados quantitativos por meio de algoritmos de aprendizado de ma¡quina existentes.
Para produzir essas descobertas, Kejriwal e Nagaraj acessaram dados por meio do corpus do Projeto Gutenberg, que contanãm 3.000 livros em inglês, uma tentativa adicional de mitigar o vianãs do pesquisador. O gaªnero de livros variou de aventura e ficção cientafica , a mistério e romance, e em meios variados, incluindo romances, contos e poesia.
Akarsh Nagaraj, MS '21, coautor do estudo e engenheiro de aprendizado de ma¡quina da Meta, ajudou a descobrir o desequilabrio litera¡rio masculino-feminino de 4:1.
“O preconceito de gaªnero émuito real, e quando vemos mulheres quatro vezes menos na literatura, isso tem um impacto subliminar nas pessoas que consomem a culturaâ€, disse Kejriwal, professor assistente de pesquisa no Departamento de Engenharia Industrial e de Sistemas Daniel J Epstein. "Revelamos quantitativamente de maneira indireta em que o preconceito persiste na cultura".
Nagaraj observou a importa¢ncia de como seus manãtodos e as descobertas do estudo lhes transmitiram uma maior compreensão dos preconceitos na sociedade e suas implicações. "Os livros são uma janela para o passado, e a escrita desses autores nos da¡ um vislumbre de como as pessoas percebem o mundo e como ele mudou."
Homens em todos os lugares… e personagens principais
O estudo descreve vários manãtodos para definir a prevalaªncia feminina na literatura. Eles utilizaram o Reconhecimento de Entidade Nomeada (NER), um manãtodo de PNL proeminente usado para extrair caracteres específicos de gaªnero. "Uma das maneiras de definir isso éolhando quantos pronomes femininos estãoem um livro em comparação com os pronomes masculinos", disse Kejriwal. A outra técnica équantificar quantas personagens femininas são os personagens principais nele.
Â
Isso permitiu que a equipe de pesquisa determinasse se os personagens masculinos eram centrais para a história.
Os achados do estudo também mostraram que a discrepa¢ncia entre personagens masculinos e femininos diminui sob autoria feminina. "Isso nos mostrou claramente que as mulheres naquela anãpoca se representavam muito mais do que um escritor masculino", disse Nagaraj.
Os manãtodos diversificados da equipe para medir e determinar a representação feminina na literatura não vieram sem limitações, no entanto, quando os autores não são nem homens nem mulheres. “Quando publicamos o artigo do conjunto de dados, os revisores tiveram essa cratica de que esta¡vamos ignorando gêneros não dicota´micosâ€, disse Kejriwal. "Mas concordamos com eles, de certa forma. Achamos que foi completamente reprimido e não conseguiremos encontrar muitos [indivaduos transgêneros ou indivíduos não dicota´micos]".
Dicotomias desafiadoras
Kejriwal reconheceu que as ferramentas de IA para identificar palavras no plural, como "eles", que podem estar se referindo a um indivaduo não dicota´mico, ainda não existem. Ainda assim, as descobertas do estudo constroem a estrutura para abordar essas questões sociais e construir as tecnologias que podem lidar com esses danãficits.
O estudo também fornece um plano para trabalhos futuros na quantificação das descobertas qualitativas que eles descobriram por meio das metodologias do estudo. Sem o vianãs inerente de pesquisas projetadas por humanos, a tecnologia da PNL também permitiu que eles encontrassem associações de adjetivos com personagens específicos de gaªnero, aprofundando sua compreensão do vianãs e sua difusão na sociedade.
"Mesmo com atribuições erradas, as palavras associadas a s mulheres eram adjetivos como 'fraca', 'ama¡vel', 'bonita' e a s vezes 'estaºpida'", disse Nagaraj. "Para personagens masculinos, as palavras que os descreviam incluaam 'liderana§a', 'poder', 'força' e 'polatica'."
Embora a equipe não tenha quantificado essa faceta de seu estudo, essa diferença nas descrições qualitativas entre caracteres específicos de gaªnero fornece um escopo futuro para uma investigação qualitativa mais abrangente sobre associações de palavras com gaªnero.
"Nosso estudo nos mostra que o mundo real écomplexo, mas hábenefacios para todos os diferentes grupos de nossa sociedade que participam do discurso cultural", disse Kejriwal. "Quando fazemos isso, tende a haver uma visão mais realista da sociedade."
Kejriwal espera que o estudo sirva para destacar a importa¢ncia da pesquisa interdisciplinar osou seja, usar a tecnologia de IA para destacar questões sociais urgentes e desigualdades que podem ser abordadas. As partes interessadas com formação especializada, incluindo cientistas da computação, podem oferecer ferramentas para processar dados e responder a perguntas, e os formuladores de políticas podem usar esses dados para promovermudanças.