The concept of the value-gradient is introduced and developed in the context of reinforcement learning, for deterministic episodic control problems that use a function approximator and have a continuous state space. It is shown that by learning the valuegradients, instead of just the values themselves, exploration or stochastic behaviour is no longer needed to find locally optimal trajectories. This is the main motivation for using value-gradients, and it is argued that learning the value-gradients is the actual objective of any value-function learning algorithm for control problems. It is also argued that learning value-gradients is significantly more efficient than learning just the values, and this argument is supported in experiments by efficiency gains of several orders of magnitude, in several problem domains. Once value-gradients are introduced into learning, several analyses become possible. For example, a surprising equivalence between a value-gradient learning algorithm and a policy-gradient learning algorithm is proven, and this provides a robust convergence proof for control problems using a value function with a general function approximator. Also, the issue of whether to include'residual gradient' terms into the weight update equations is addressed. Finally, an analysis is made of actor-critic architectures, which finds strong similarities to back-propagation through time, and gives simplifications and convergence proofs to certain actor-critic architectures, but while making those actor-critic architectures redundant. Unfortunately, by proving equivalence to policy-gradient learning, finding new divergence examples even in the absence of bootstrapping, and proving the redundancy of residual-gradients and actor-critic architectures in some circumstances, this paper does somewhat discredit the usefulness of using a value-function.

Stress and genetic background regulate different aspects of behavioral learning through the action of stress hormones and neuromodulators.

Autonomous helicopter flight is widely regarded to be a highly challenging control problem. This paper presents the first successful autonomous completion on a real RC helicopter of the following four aerobatic maneuvers: forward flip and sideways roll at low speed, tail-in funnel, and nose-in funnel. Our experimental results significantly extend the state of the art in autonomous helicopter flight. We used the following approach: First we had a pilot fly the helicopter to help us find a helicopter dynamics model and a reward (cost) function. Then we used a reinforcement learning (optimal control) algorithm to find a controller that is optimized for the resulting model and reward function. More specifically, we used differential dynamic programming (DDP), an extension of the linear quadratic regulator (LQR).

We introduce a class of MPDs which greatly simplify Reinforcement Learning. They have discrete state spaces and continuous control spaces. The controls have the effect of rescaling the transition probabilities of an underlying Markov chain. A control cost penalizing KL divergence between controlled and uncontrolled transition probabilities makes the minimization problem convex, and allows analytical computation of the optimal controls given the optimal value function. An exponential transformation of the optimal value function makes the minimized Bellman equation linear.

Current road-traffic optimisation practice around the world is a combination of hand tuned policies with a small degree of automatic adaption. Even state-ofthe-art research controllers need good models of the road traffic, which cannot be obtained directly from existing sensors. We use a policy-gradient reinforcement learning approach to directly optimise the traffic signals, mapping currently deployed sensor observations to control signals. Our trained controllers are (theoretically) compatible with the traffic system used in Sydney and many other cities around the world. We apply two policy-gradient methods: (1) the recent natural actor-critic algorithm, and (2) a vanilla policy-gradient algorithm for comparison. Along the way we extend natural-actor critic approaches to work for distributed and online infinite-horizon problems.

Policy gradient methods are reinforcement learning algorithms that adapt a parameterized policy by following a performance gradient estimate. Conventional policy gradient methods use Monte-Carlo techniques to estimate this gradient. Since Monte Carlo methods tend to have high variance, a large number of samples is required, resulting in slow convergence. In this paper, we propose a Bayesian framework that models the policy gradient as a Gaussian process. This reduces the number of samples needed to obtain accurate gradient estimates. Moreover, estimates of the natural gradient as well as a measure of the uncertainty in the gradient estimates are provided at little extra cost.

In this paper, we generalize the previous iLSTD algorithm and present three new results: (1) the first convergence proof for an iLSTD algorithm; (2) an extension to incorporate eligibility traces without changing the asymptotic computational complexity; and (3) the first empirical results with an iLSTD algorithm for a problem (mountain car) with feature vectors large enough (n 10, 000) to show substantial computational advantages over LSTD.

We present a learning algorithm for undiscounted reinforcement learning. Our interest lies in bounds for the algorithm's online performance after some finite number of steps. In the spirit of similar methods already successfully applied for the exploration-exploitation tradeoff in multi-armed bandit problems, we use upper confidence bounds to show that our UCRL algorithm achieves logarithmic online regret in the number of steps taken with respect to an optimal policy.

We introduce a class of MPDs which greatly simplify Reinforcement Learning. They have discrete state spaces and continuous control spaces. The controls have the effect of rescaling the transition probabilities of an underlying Markov chain. A control cost penalizing KL divergence between controlled and uncontrolled transition probabilities makes the minimization problem convex, and allows analytical computation of the optimal controls given the optimal value function. An exponential transformation of the optimal value function makes the minimized Bellman equation linear.

Current road-traffic optimisation practice around the world is a combination of hand tuned policies with a small degree of automatic adaption. Even state-ofthe-art research controllers need good models of the road traffic, which cannot be obtained directly from existing sensors. We use a policy-gradient reinforcement learning approach to directly optimise the traffic signals, mapping currently deployed sensor observations to control signals. Our trained controllers are (theoretically) compatible with the traffic system used in Sydney and many other cities around the world. We apply two policy-gradient methods: (1) the recent natural actor-critic algorithm, and (2) a vanilla policy-gradient algorithm for comparison. Along the way we extend natural-actor critic approaches to work for distributed and online infinite-horizon problems.