Intelligent support for Human Oversight: Integrating Reinforcement Learning with Gaze Simulation to Personalize Highlighting
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866918328682938368 |
|---|---|
| author | Klößner, Thorsten Belo, João Wu, Zekun Hoffmann, Jörg Feit, Anna Maria |
| author_facet | Klößner, Thorsten Belo, João Wu, Zekun Hoffmann, Jörg Feit, Anna Maria |
| contents | Interfaces for human oversight must effectively support users' situation awareness under time-critical conditions. We explore reinforcement learning (RL)-based UI adaptation to personalize alerting strategies that balance the benefits of highlighting critical events against the cognitive costs of interruptions. To enable learning without real-world deployment, we integrate models of users' gaze behavior to simulate attentional dynamics during monitoring. Using a delivery-drone oversight scenario, we present initial results suggesting that RL-based highlighting can outperform static, rule-based approaches and discuss challenges of intelligent oversight support. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_08403 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Intelligent support for Human Oversight: Integrating Reinforcement Learning with Gaze Simulation to Personalize Highlighting Klößner, Thorsten Belo, João Wu, Zekun Hoffmann, Jörg Feit, Anna Maria Human-Computer Interaction Artificial Intelligence Interfaces for human oversight must effectively support users' situation awareness under time-critical conditions. We explore reinforcement learning (RL)-based UI adaptation to personalize alerting strategies that balance the benefits of highlighting critical events against the cognitive costs of interruptions. To enable learning without real-world deployment, we integrate models of users' gaze behavior to simulate attentional dynamics during monitoring. Using a delivery-drone oversight scenario, we present initial results suggesting that RL-based highlighting can outperform static, rule-based approaches and discuss challenges of intelligent oversight support. |
| title | Intelligent support for Human Oversight: Integrating Reinforcement Learning with Gaze Simulation to Personalize Highlighting |
| topic | Human-Computer Interaction Artificial Intelligence |
| url | https://arxiv.org/abs/2602.08403 |