Tag Archives: Exploration Vs. Exploitation

Finding the common utility of actions in several tasks learnt in the same domain in order to reduce the learning cost of reinforcement learning

Rosman, B.; Ramamoorthy, S., Action Priors for Learning Domain Invariances, Autonomous Mental Development, IEEE Transactions on , vol.7, no.2, pp.107,118, June 2015, DOI: 10.1109/TAMD.2015.2419715.

An agent tasked with solving a number of different decision making problems in similar environments has an opportunity to learn over a longer timescale than each individual task. Through examining solutions to different tasks, it can uncover behavioral invariances in the domain, by identifying actions to be prioritized in local contexts, invariant to task details. This information has the effect of greatly increasing the speed of solving new problems. We formalise this notion as action priors, defined as distributions over the action space, conditioned on environment state, and show how these can be learnt from a set of value functions. We apply action priors in the setting of reinforcement learning, to bias action selection during exploration. Aggressive use of action priors performs context based pruning of the available actions, thus reducing the complexity of lookahead during search. We additionally define action priors over observation features, rather than states, which provides further flexibility and generalizability, with the additional benefit of enabling feature selection. Action priors are demonstrated in experiments in a simulated factory environment and a large random graph domain, and show significant speed ups in learning new tasks. Furthermore, we argue that this mechanism is cognitively plausible, and is compatible with findings from cognitive psychology.

Efficient sampling of the agent-world interaction in reinforcement learning through the use of simulators with diverse fidelity to the real system

Cutler, M.; Walsh, T.J.; How, J.P., Real-World Reinforcement Learning via Multifidelity Simulators, Robotics, IEEE Transactions on , vol.31, no.3, pp.655,671, June 2015, DOI: 10.1109/TRO.2015.2419431.

Reinforcement learning (RL) can be a tool for designing policies and controllers for robotic systems. However, the cost of real-world samples remains prohibitive as many RL algorithms require a large number of samples before learning useful policies. Simulators are one way to decrease the number of required real-world samples, but imperfect models make deciding when and how to trust samples from a simulator difficult. We present a framework for efficient RL in a scenario where multiple simulators of a target task are available, each with varying levels of fidelity. The framework is designed to limit the number of samples used in each successively higher-fidelity/cost simulator by allowing a learning agent to choose to run trajectories at the lowest level simulator that will still provide it with useful information. Theoretical proofs of the framework’s sample complexity are given and empirical results are demonstrated on a remote-controlled car with multiple simulators. The approach enables RL algorithms to find near-optimal policies in a physical robot domain with fewer expensive real-world samples than previous transfer approaches or learning without simulators.

Neurological evidences of the hierarchical arrangement of the process of motor skill learning

Jörn Diedrichsen, Katja Kornysheva, Motor skill learning between selection and execution, Trends in Cognitive Sciences, Volume 19, Issue 4, April 2015, Pages 227-233, ISSN 1364-6613, DOI: 10.1016/j.tics.2015.02.003.

Learning motor skills evolves from the effortful selection of single movement elements to their combined fast and accurate production. We review recent trends in the study of skill learning which suggest a hierarchical organization of the representations that underlie such expert performance, with premotor areas encoding short sequential movement elements (chunks) or particular component features (timing/spatial organization). This hierarchical representation allows the system to utilize elements of well-learned skills in a flexible manner. One neural correlate of skill development is the emergence of specialized neural circuits that can produce the required elements in a stable and invariant fashion. We discuss the challenges in detecting these changes with fMRI.

Active exploration strategy for RL in robots, and approximation of value function by Gaussian processes

Jen Jen Chung, Nicholas R.J. Lawrance, Salah Sukkarieh (2015), Learning to soar: Resource-constrained exploration in reinforcement learning, The International Journal of Robotics Research vol. 34, pp. 158-172. DOI: 10.1177/0278364914553683

This paper examines temporal difference reinforcement learning with adaptive and directed exploration for resource-limited missions. The scenario considered is that of an unpowered aerial glider learning to perform energy-gaining flight trajectories in a thermal updraft. The presented algorithm, eGP-SARSA(\u03bb), uses a Gaussian process regression model to estimate the value function in a reinforcement learning framework. The Gaussian process also provides a variance on these estimates that is used to measure the contribution of future observations to the Gaussian process value function model in terms of information gain. To avoid myopic exploration we developed a resource-weighted objective function that combines an estimate of the future information gain using an action rollout with the estimated value function to generate directed explorative action sequences. A number of modifications and computational speed-ups to the algorithm are presented along with a standard GP-SARSA(\u03bb) implementation with Formula -greedy exploration to compare the respective learning performances. The results show that under this objective function, the learning agent is able to continue exploring for better state-action trajectories when platform energy is high and follow conservative energy-gaining trajectories when platform energy is low.

On search as a consequence of the exploration-exploitation trade-off, and as a core element in human cognition

Thomas T. Hills, Peter M. Todd, David Lazer, A. David Redish, Iain D. Couzin, the Cognitive Search Research Group, Exploration versus exploitation in space, mind, and society, Trends in Cognitive Sciences, Volume 19, Issue 1, January 2015, Pages 46-54, ISSN 1364-6613, DOI: 10.1016/j.tics.2014.10.004.

Search is a ubiquitous property of life. Although diverse domains have worked on search problems largely in isolation, recent trends across disciplines indicate that the formal properties of these problems share similar structures and, often, similar solutions. Moreover, internal search (e.g., memory search) shows similar characteristics to external search (e.g., spatial foraging), including shared neural mechanisms consistent with a common evolutionary origin across species. Search problems and their solutions also scale from individuals to societies, underlying and constraining problem solving, memory, information search, and scientific and cultural innovation. In summary, search represents a core feature of cognition, with a vast influence on its evolution and processes across contexts and requiring input from multiple domains to understand its implications and scope.

A good summary and classification of state-of-the-art motion planning algorithms and proposal of a new one that improve the expected computational cost

Rickert, M.; Sieverling, A.; Brock, O., Balancing Exploration and Exploitation in Sampling-Based Motion Planning, Robotics, IEEE Transactions on , vol.30, no.6, pp.1305,1317, Dec. 2014. DOI: 10.1109/TRO.2014.2340191

We present the exploring/exploiting tree (EET) algorithm for motion planning. The EET planner deliberately trades probabilistic completeness for computational efficiency. This tradeoff enables the EET planner to outperform state-of-the-art sampling-based planners by up to three orders of magnitude. We show that these considerable speedups apply for a variety of challenging real-world motion planning problems. The performance improvements are achieved by leveraging work space information to continuously adjust the sampling behavior of the planner. When the available information captures the planning problem’s inherent structure, the planner’s sampler becomes increasingly exploitative. When the available information is less accurate, the planner automatically compensates by increasing local configuration space exploration. We show that active balancing of exploration and exploitation based on workspace information can be a key ingredient to enabling highly efficient motion planning in practical scenarios.