| Budget Amount *help |
¥43,160,000 (Direct Cost: ¥33,200,000、Indirect Cost: ¥9,960,000)
Fiscal Year 2019: ¥10,920,000 (Direct Cost: ¥8,400,000、Indirect Cost: ¥2,520,000)
Fiscal Year 2018: ¥11,050,000 (Direct Cost: ¥8,500,000、Indirect Cost: ¥2,550,000)
Fiscal Year 2017: ¥10,530,000 (Direct Cost: ¥8,100,000、Indirect Cost: ¥2,430,000)
Fiscal Year 2016: ¥10,660,000 (Direct Cost: ¥8,200,000、Indirect Cost: ¥2,460,000)
|
|---|
| Outline of Final Research Achievements |
The key parameters in the computational theory of reinforcement learning are the learning rate α, which determines the speed of learning, and the inverse temperature β, which determines the balance between "use of previously available information and search for new information." We tested the hypothesis that the cognitive mechanism characterized by these two parameters can be an explanatory principle not only for reinforcement learning itself and decision making, but also for problem solving. As a result of our experiments, we found a significant correlation between the inverse temperature β measured in the bandit task and the creativity/uniqueness scores measured in the UUT task, one of the idea generation tasks. Thus, the strength of the tendency to search for novel information in reinforcement learning could explain the creativity in the idea generation task.
|
