Scholarlink

Cancer research is a vital area that has been the subject of intense research, and with the increasing use of AI in the field, there is a growing need for transparency and explainability. However, this also has to answer the needs of specific users and domain related challenges. This paper presents a preliminary study under the INCISIVE project that designs an eXplainable AI (XAI) questionnaire addressing Health Professionals’ explainability needs in AI for cancer imaging research. The questionnaire, developed with principles of transparency and user-centered design, incorporates perspectives and requirements of medical experts. Results from applying the questionnaire to AI cancer imaging researchers demonstrate its potential as a transparency-promoting tool, despite certain limitations.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Designing a Method to Identify Explainability Requirements in Cancer Research

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Didier Dominguez,
  • Dimitris Fotopoulos,
  • Ioanna Chouvarda,
  • Susanna Ausso

摘要

Cancer research is a vital area that has been the subject of intense research, and with the increasing use of AI in the field, there is a growing need for transparency and explainability. However, this also has to answer the needs of specific users and domain related challenges. This paper presents a preliminary study under the INCISIVE project that designs an eXplainable AI (XAI) questionnaire addressing Health Professionals’ explainability needs in AI for cancer imaging research. The questionnaire, developed with principles of transparency and user-centered design, incorporates perspectives and requirements of medical experts. Results from applying the questionnaire to AI cancer imaging researchers demonstrate its potential as a transparency-promoting tool, despite certain limitations.