Object-oriented representations in reinforcement learning have shown promise in transfer learning, with previous research introducing a propositional object-oriented framework that has provably efficient learning bounds with respect to sample complexity. However, this framework has limitations in terms of the classes of tasks it can efficiently learn. In this paper we introduce a novel deictic object-oriented framework that has provably efficient learning bounds and can solve a broader range of tasks. Additionally, we show that this framework is capable of zero-shot transfer of transition dynamics across tasks and demonstrate this empirically for the Taxi and Sokoban domains.

Object-oriented representations in reinforcement learning have shown promise in transfer learning, with previous research introducing a propositional object-oriented framework that has provably efficient learning bounds with respect to sample complexity. However, this framework has limitations in terms of the classes of tasks it can efficiently learn. In this paper we introduce a novel deictic object-oriented framework that has provably efficient learning bounds and can solve a broader range of tasks. Additionally, we show that this framework is capable of zero-shot transfer of transition dynamics across tasks and demonstrate this empirically for the Taxi and Sokoban domains.

The field of software simulators for neural networks has been ex(cid:173) panding very rapidly in the last years but their importance is still being underestimated. They must provide increasing levels of as(cid:173) sistance for the design, simulation and analysis of neural networks. With our object-oriented framework (SESAME) we intend to show that very high degrees of transparency, manageability and flexibil(cid:173) ity for complex experiments can be obtained. SESAME's basic de(cid:173) sign philosophy is inspired by the natural way in which researchers explain their computational models. Experiments are performed with networks of building blocks, which can be extended very eas(cid:173) ily.

Object-oriented representations in reinforcement learning have shown promise in transfer learning, with previous research introducing a propositional object-oriented framework that has provably efficient learning bounds with respect to sample complexity. However, this framework has limitations in terms of the classes of tasks it can efficiently learn. In this paper we introduce a novel deictic object-oriented framework that has provably efficient learning bounds and can solve a broader range of tasks. Additionally, we show that this framework is capable of zero-shot transfer of transition dynamics across tasks and demonstrate this empirically for the Taxi and Sokoban domains. Papers published at the Neural Information Processing Systems Conference.

The field of software simulators for neural networks has been expanding very rapidly in the last years but their importance is still being underestimated. They must provide increasing levels of assistance for the design, simulation and analysis of neural networks. With our object-oriented framework (SESAME) we intend to show that very high degrees of transparency, manageability and flexibility for complex experiments can be obtained. SESAME's basic design philosophy is inspired by the natural way in which researchers explain their computational models. Experiments are performed with networks of building blocks, which can be extended very easily.

The field of software simulators for neural networks has been expanding very rapidly in the last years but their importance is still being underestimated. They must provide increasing levels of assistance for the design, simulation and analysis of neural networks. With our object-oriented framework (SESAME) we intend to show that very high degrees of transparency, manageability and flexibility for complex experiments can be obtained. SESAME's basic design philosophy is inspired by the natural way in which researchers explain their computational models. Experiments are performed with networks of building blocks, which can be extended very easily.

The field of software simulators for neural networks has been expanding veryrapidly in the last years but their importance is still being underestimated. They must provide increasing levels of assistance forthe design, simulation and analysis of neural networks. With our object-oriented framework (SESAME) we intend to show that very high degrees of transparency, manageability and flexibility forcomplex experiments can be obtained. SESAME's basic design philosophyis inspired by the natural way in which researchers explain their computational models. Experiments are performed with networks of building blocks, which can be extended very easily. Mechanismshave been integrated to facilitate the construction and analysis of very complex architectures.