imitation gap
On Imitation in Mean-field Games
In this paper, departing from the existing literature on IL for MFGs, we introduce a new solution concept called the Nash imitation gap.
- Asia > China > Shanghai > Shanghai (0.04)
- Europe > Switzerland > Zürich > Zürich (0.04)
- Europe > Germany > Baden-Württemberg > Tübingen Region > Tübingen (0.04)
On Imitation in Mean-field Games
In this paper, departing from the existing literature on IL for MFGs, we introduce a new solution concept called the Nash imitation gap.
- Asia > China > Shanghai > Shanghai (0.04)
- Europe > Switzerland > Zürich > Zürich (0.04)
- Europe > Germany > Baden-Württemberg > Tübingen Region > Tübingen (0.04)
- Information Technology > Artificial Intelligence > Machine Learning > Reinforcement Learning (1.00)
- Information Technology > Artificial Intelligence > Representation & Reasoning > Agents (0.94)
- Information Technology > Artificial Intelligence > Robots (0.71)
- Information Technology > Artificial Intelligence > Machine Learning > Neural Networks (0.68)
- North America > United States (0.04)
- Asia > China > Hong Kong (0.04)
- Asia > China > Guangdong Province > Shenzhen (0.04)
- Asia > China > Jiangsu Province > Nanjing (0.04)
- Information Technology (0.67)
- Leisure & Entertainment > Games > Computer Games (0.47)
Bridging the Imitation Gap by Adaptive Insubordination
In practice, imitation learning is preferred over pure reinforcement learning whenever it is possible to design a teaching agent to provide expert supervision. However, we show that when the teaching agent makes decisions with access to privileged information that is unavailable to the student, this information is marginalized during imitation learning, resulting in an imitation gap and, potentially, poor results.
- Asia > China > Shanghai > Shanghai (0.04)
- Europe > Switzerland > Zürich > Zürich (0.04)
- Europe > Germany > Baden-Württemberg > Tübingen Region > Tübingen (0.04)
On Imitation in Mean-field Games
In this paper, departing from the existing literature on IL for MFGs, we introduce a new solution concept called the Nash imitation gap.
- Asia > China > Shanghai > Shanghai (0.04)
- Europe > Switzerland > Zürich > Zürich (0.04)
- Europe > Germany > Baden-Württemberg > Tübingen Region > Tübingen (0.04)
- Information Technology > Artificial Intelligence > Machine Learning > Reinforcement Learning (1.00)
- Information Technology > Artificial Intelligence > Representation & Reasoning > Agents (0.94)
- Information Technology > Artificial Intelligence > Robots (0.71)
- Information Technology > Artificial Intelligence > Machine Learning > Neural Networks (0.68)
A Broader Impacts
The last inequality follows Eq. (18).
- North America > United States (0.04)
- Asia > China > Hong Kong (0.04)
- Asia > China > Guangdong Province > Shenzhen (0.04)
- Asia > China > Jiangsu Province > Nanjing (0.04)
- Information Technology (0.67)
- Leisure & Entertainment > Games > Computer Games (0.47)