Reinforcement learning, which acquires a policy maximizing long-term rewards, has been actively studied. Unfortunately, this learning type is too slow and difficult to use in practical situations because the state-action space becomes huge in real environments. Many studies have incorporated human knowledge into reinforcement Learning. Though human knowledge on trajectories is often used, a human could be asked to control an AI agent, which can be difficult. Knowledge on subgoals may lessen this requirement because humans need only to consider a few representative states on an optimal trajectory in their minds. The essential factor for learning efficiency is rewards. Potential-based reward shaping is a basic method for enriching rewards. However, it is often difficult to incorporate subgoals for accelerating learning over potential-based reward shaping. This is because the appropriate potentials are not intuitive for humans. We extend potential-based reward shaping and propose a subgoal-based reward shaping. The method makes it easier for human trainers to share their knowledge of subgoals. To evaluate our method, we obtained a subgoal series from participants and conducted experiments in three domains, fourrooms(discrete states and discrete actions), pinball(continuous and discrete), and picking(both continuous). We compared our method with a baseline reinforcement learning algorithm and other subgoal-based methods, including random subgoal and naive subgoal-based reward shaping. As a result, we found out that our reward shaping outperformed all other methods in learning efficiency.

We propose a method for measuring interdisciplinary research by dividing it into two approaches: interdisciplinary research conducted by individual researchers and interdisciplinary research involving the collaboration of multiple researchers. Using this method, a database of “KAKENHI,” which is a grant-in-aid for scientific research provided by the Japan Society for the Promotion of Science (JSPS), is employed to measure interdisciplinary research from the perspective of the two research approaches, and the features of interdisciplinary research in KAKENHI are analyzed. The analysis results indicate the following: (1) the number of collaborative interdisciplinary research projects is larger than the number of individual interdisciplinary research projects, (2) the number of interdisciplinary research projects for each field and for each combination of fields differs among fields, and (3) the relationship between the numbers of interdisciplinary research projects in the two fields is asymmetric with regard to the main- and sub-fields of interdisciplinary research. As the proposed measurement method is capable of quantitatively measuring interdisciplinarity between fields and their research organizations, it will be useful for decision-makers in science and technology policy and strategy.

No abstract

No abstract