Learn from user feedback

AI Model
Neural Network Probabilistic Unknown White Box
AI Task
Multi-class Classification Regression Text Classification
Application Domain
Finance/Economics Media and Communication Recommendation
Type of Users
Domain-experts Generic Non-experts
Explanation Modality
Natural Language Text Visual
XAI Model
Counter-exemplars/factual Exemplars Features Importance None
Related Papers

The system should learn from the information the user inputs into it [10.1016/j.artint.2021.103507].

User feedback design plays a key role in retaining users during the problem-solving process. Conversational user interfaces are needed to integrate more advanced human-AI interaction, natural language processing, machine learning and pedagogical design strategies to satisfy users’ needs [10.1007/978-3-031-17615-9_34].

Explanations can be used for calibrating trust, improving the user’s task skills, collaborating more effectively with AI, and giving constructive feedback to developers [10.1145/3544548.3581001].

The ability to provide high-level feedback can, in fact, significantly improves the user’s sense of trust and control and system transparency [10.1145/3589345].

The feedback to the system should be [10.1145/3514258]: 1) Actionable 2) Reversible 3) Honored 4) Shown to incremental changes