Scholarlink

The central topic concerning this work was fully automatizing the creation of causal explanations for model-free RL agents. The resulting textual answers to “Why” and “Why not” questions had to maintain a high level of understandability in order to improve the trust of users in the algorithmic choices. To achieve this goal, a complex ensemble methodology has been developed: Auto-BENEDICT.

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

Conclusion and Future Work

  • 章节

  • 发表日期：

    2026年

查看全文
  • Rudy Milani

摘要

The central topic concerning this work was fully automatizing the creation of causal explanations for model-free RL agents. The resulting textual answers to “Why” and “Why not” questions had to maintain a high level of understandability in order to improve the trust of users in the algorithmic choices. To achieve this goal, a complex ensemble methodology has been developed: Auto-BENEDICT.