When a new-born animal (agent) opens its eyes, what it sees is a patchwork of light and dark patterns, the natural scene.What is perceived by the agent at this moment is based on the patternof neural spikes in its brain. Life-long learning begins with such a flood of spikes in the brain. All knowledge and skills learned by the agent are mediated by such spikes, thus it is critical to understand what information these spikes convey and how they can be used to generate meaningful behavior. Here, we consider how agents can autonomously understand the meaning of these spikes without direct reference to the stimulus. We find that this problem, the problem of grounding, is unsolvable if the agent is passively perceiving, and that it can be solved only through self-initiated action. Furthermore, we show that a simple criterion, combined with standard reinforcement learning, can help solve this problem. We will present simulation results and discuss the implications of these results on life-long learning.

We describe a continuous state/action reinforcement learning method which uses deep belief networks (DBNs) in conjunction with a value function-based reinforcement learning algorithm to learn effective control policies. Our approach is to first learn a model of the state-action space from data in an unsupervised pre-training phase, and then use neural-fitted Q-iteration (NFQ) to learn an accurate value function approximator (analogous to a "fine-tuning" phase when training DBNs for classification). Our experiments suggest that this approach has the potential to significantly increase the efficiency of the learning process in NFQ, provided care is taken to ensure the initial data covers interesting areas of the state-action space, and may be particularly useful in transfer learning settings.

Integrating games into the computer science curriculum has been gaining acceptance in recent years, particularly when used to improve student engagement in introductory courses. This paper argues that games can also be useful in upper level courses, such as general artificial intelligence and machine learning. We provide a case study of using a Mario game in a machine learning class to provide one successful data point where both content-specific and general learning outcomes were successfully achieved.

In this work we present an advanced Bayesian formulation to the task of control learning that employs the Relevance Vector Machines (RVM) generative model for value function evaluation. The key aspect of the proposed method is the design of the discount return as a generalized linear model that constitutes a well-known probabilistic approach. This allows to augment the model with advantageous sparse priors provided by the RVM's regression framework. We have also taken into account the significant issue of selecting the proper parameters of the kernel design matrix. Experiments have shown that our method produces improved performance in both simulated and real test environments.

We consider settings in which a semi-autonomous agent has uncertain knowledge about its environment, but can ask what action the human operator would prefer taking in the current or in a potential future state. Asking queries can improve behavior, but if queries come at a cost (e.g., due to limited operator attention), the value of each query should be maximized. We compare two strategies for selecting action queries: 1) based on myopically maximizing expected gain in long-term value, and 2) based on myopically minimizing uncertainty in the agent's policy representation. We show empirically that the first strategy tends to select more valuable queries, and that a hybrid method can outperform either method alone in settings with limited computation.

In many multi-agent applications such as distributed sensor nets, a network of agents act collaboratively under uncertainty and local interactions. Networked Distributed POMDP (ND-POMDP) provides a framework to model such cooperative multi-agent decision making. Existing work on ND-POMDPs has focused on offline techniques that require accurate models, which are usually costly to obtain in practice. This paper presents a model-free, scalable learning approach that synthesizes multi-agent reinforcement learning (MARL) and distributed constraint optimization (DCOP). By exploiting structured interaction in ND-POMDPs, our approach distributes the learning of the joint policy and employs DCOP techniques to coordinate distributed learning to ensure the global learning performance. Our approach can learn a globally optimal policy for ND-POMDPs with a property called groupwise observability. Experimental results show that, with communication during learning and execution, our approach significantly outperforms the nearly-optimal non-communication policies computed offline.

In this paper we propose differential eligibility vectors (DEV) for temporal-difference (TD) learning, a new class of eligibility vectors designed to bring out the contribution of each action in the TD-error at each state. Specifically, we use DEV in TD-Q(lambda) to more accurately learn the relative value of the actions, rather than their absolute value. We identify conditions that ensure convergence w.p.1 of TD-Q(lambda) with DEV and show that this algorithm can also be used to directly approximate the advantage function associated with a given policy, without the need to compute an auxiliary function - something that, to the extent of our knowledge, was not known possible. Finally, we discuss the integration of DEV in LSTDQ and actor-critic algorithms.

Recent Reinforcement Learning (RL) algorithms, such as R-MAX, make (with high probability) only a small number of poor decisions. In practice, these algorithms do not scale well as the number of states grows because the algorithms spend too much effort exploring. We introduce an RL algorithm State TArgeted R-MAX (STAR-MAX) that explores a subset of the state space, called the exploration envelope ξ. When ξ equals the total state space, STAR-MAX behaves identically to R-MAX. When ξ is a subset of the state space, to keep exploration within ξ, a recovery rule β is needed. We compared existing algorithms with our algorithm employing various exploration envelopes. With an appropriate choice of ξ, STAR-MAX scales far better than existing RL algorithms as the number of states increases. A possible drawback of our algorithm is its dependence on a good choice of ξ and β. However, we show that an effective recovery rule β can be learned on-line and ξ can be learned from demonstrations. We also find that even randomly sampled exploration envelopes can improve cumulative rewards compared to R-MAX. We expect these results to lead to more efficient methods for RL in large-scale problems.

We describe the Fourier basis, a linear value function approximation scheme based on the Fourier series. We empirically demonstrate that it performs well compared to radial basis functions and the polynomial basis, the two most popular fixed bases for linear value function approximation, and is competitive with learned proto-value functions.

We study the problem of automatically generating features for function approximation in reinforcement learning. We build on the work of Mahadevan and his colleagues, who pioneered the use of spectral clustering methods for basis function construction. Their methods work on top of a graph that captures state adjacency. Instead, we use bisimulation metrics in order to provide state distances for spectral clustering. The advantage of these metrics is that they incorporate reward information in a natural way, in addition to the state transition information. We provide theoretical bounds on the quality of the obtained approximation, which justify the importance of incorporating reward information. We also demonstrate empirically that the approximation quality improves when bisimulation metrics are used instead of the state adjacency graph in the basis function construction process.