Tag Archives: Reinforcement Learning

Reinforcement learning applied to select which parts of a Neural Turing Machine are to be updated with backpropagation during learning

May 12, 2015 18:21 ,

Wojciech Zaremba, Ilya Sutskever, Reinforcement Learning Neural Turing Machines, arXiv.org, arXiv:1505.00521.

The expressive power of a machine learning model is closely related to the number of sequential computational steps it can learn. For example, Deep Neural Networks have been more successful than shallow networks because they can perform a greater number of sequential computational steps (each highly parallel). The Neural Turing Machine (NTM) is a model that can compactly express an even greater number of sequential computational steps, so it is even more powerful than a DNN. Its memory addressing operations are designed to be differentiable; thus the NTM can be trained with backpropagation.
While differentiable memory is relatively easy to implement and train, it necessitates accessing the entire memory content at each computational step. This makes it difficult to implement a fast NTM. In this work, we use the Reinforce algorithm to learn where to access the memory, while using backpropagation to learn what to write to the memory. We call this model the RL-NTM. Reinforce allows our model to access a constant number of memory cells at each computational step, so its implementation can be faster. The RL-NTM is the first model that can, in principle, learn programs of unbounded running time. We successfully trained the RL-NTM to solve a number of algorithmic tasks that are simpler than the ones solvable by the fully differentiable NTM.
As the RL-NTM is a fairly intricate model, we needed a method for verifying the correctness of our implementation. To do so, we developed a simple technique for numerically checking arbitrary implementations of models that use Reinforce, which may be of independent interest.

Posted in: Cognitive sciences , Tagged: , ,

Reinforcement learning to explain emotions

May 4, 2015 11:59 ,

Joost Broekensa, Elmer Jacobsa & Catholijn M. Jonker, A reinforcement learning model of joy, distress, hope and fear, Connection Science, DOI: 10.1080/09540091.2015.1031081.

In this paper we computationally study the relation between adaptive behaviour and emotion. Using the reinforcement learning framework, we propose that learned state utility, V(s), models fear (negative) and hope (positive) based on the fact that both signals are about anticipation of loss or gain. Further, we propose that joy/distress is a signal similar to the error signal. We present agent-based simulation experiments that show that this model replicates psychological and behavioural dynamics of emotion. This work distinguishes itself by assessing the dynamics of emotion in an adaptive agent framework – coupling it to the literature on habituation, development, extinction and hope theory. Our results support the idea that the function of emotion is to provide a complex feedback signal for an organism to adapt its behaviour. Our work is relevant for understanding the relation between emotion and adaptation in animals, as well as for human–robot interaction, in particular how emotional signals can be used to communicate between adaptive agents and humans.

Posted in: Psycho-physiological bases of engineering , Tagged: ,

Application of reinforcement learning to the defense against attacks on communication networks

May 4, 2015 11:31 ,

Kleanthis Malialisa, Sam Devlina & Daniel Kudenkoa, Distributed reinforcement learning for adaptive and robust network intrusion response, Connection Science, DOI: 0.1080/09540091.2015.1031082.

Distributed denial of service (DDoS) attacks constitute a rapidly evolving threat in the current Internet. Multiagent Router Throttling is a novel approach to defend against DDoS attacks where multiple reinforcement learning agents are installed on a set of routers and learn to rate-limit or throttle traffic towards a victim server. The focus of this paper is on online learning and scalability. We propose an approach that incorporates task decomposition, team rewards and a form of reward shaping called difference rewards. One of the novel characteristics of the proposed system is that it provides a decentralised coordinated response to the DDoS problem, thus being resilient to DDoS attacks themselves. The proposed system learns remarkably fast, thus being suitable for online learning. Furthermore, its scalability is successfully demonstrated in experiments involving 1000 learning agents. We compare our approach against a baseline and a popular state-of-the-art throttling technique from the network security literature and show that the proposed approach is more effective, adaptive to sophisticated attack rate dynamics and robust to agent failures.

Posted in: Communication networks , Tagged:

Reinforcement learning used for an adaptive attention mechanism, and integrated in an architecture with both top-down and bottom-up vision processing

April 24, 2015 11:26 ,

Ognibene, D.; Baldassare, G., Ecological Active Vision: Four Bioinspired Principles to Integrate Bottom–Up and Adaptive Top–Down Attention Tested With a Simple Camera-Arm Robot, Autonomous Mental Development, IEEE Transactions on , vol.7, no.1, pp.3,25, March 2015. DOI: 10.1109/TAMD.2014.2341351.

Vision gives primates a wealth of information useful to manipulate the environment, but at the same time it can easily overwhelm their computational resources. Active vision is a key solution found by nature to solve this problem: a limited fovea actively displaced in space to collect only relevant information. Here we highlight that in ecological conditions this solution encounters four problems: 1) the agent needs to learn where to look based on its goals; 2) manipulation causes learning feedback in areas of space possibly outside the attention focus; 3) good visual actions are needed to guide manipulation actions, but only these can generate learning feedback; and 4) a limited fovea causes aliasing problems. We then propose a computational architecture (“BITPIC”) to overcome the four problems, integrating four bioinspired key ingredients: 1) reinforcement-learning fovea-based top-down attention; 2) a strong vision-manipulation coupling; 3) bottom-up periphery-based attention; and 4) a novel action-oriented memory. The system is tested with a simple simulated camera-arm robot solving a class of search-and-reach tasks involving color-blob “objects.” The results show that the architecture solves the problems, and hence the tasks, very efficiently, and highlight how the architecture principles can contribute to a full exploitation of the advantages of active vision in ecological conditions.

Posted in: Applications of reinforcement learning to robots, Computer vision, Psycho-physiological bases of engineering , Tagged: , , , ,

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

February 10, 2015 16:49 ,

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.

Posted in: Applications of reinforcement learning to robots , Tagged: , ,

Solving the problem of the slow learning rate of reinfocerment learning through the acquisition of the transition model from the data

January 23, 2015 18:03 ,

Deisenroth, M.P.; Fox, D.; Rasmussen, C.E., Gaussian Processes for Data-Efficient Learning in Robotics and Control, Pattern Analysis and Machine Intelligence, IEEE Transactions on , vol.37, no.2, pp.408,423, Feb. 2015, DOI: 10.1109/TPAMI.2013.218

Autonomous learning has been a promising direction in control and robotics for more than a decade since data-driven learning allows to reduce the amount of engineering knowledge, which is otherwise required. However, autonomous reinforcement learning (RL) approaches typically require many interactions with the system to learn controllers, which is a practical limitation in real systems, such as robots, where many interactions can be impractical and time consuming. To address this problem, current learning approaches typically require task-specific knowledge in form of expert demonstrations, realistic simulators, pre-shaped policies, or specific knowledge about the underlying dynamics. In this paper, we follow a different approach and speed up learning by extracting more information from data. In particular, we learn a probabilistic, non-parametric Gaussian process transition model of the system. By explicitly incorporating model uncertainty into long-term planning and controller learning our approach reduces the effects of model errors, a key problem in model-based learning. Compared to state-of-the art RL our model-based policy search method achieves an unprecedented speed of learning. We demonstrate its applicability to autonomous learning in real robot and control tasks.

Posted in: Applications of reinforcement learning to robots , Tagged: ,

Partially observable reinforcement learning and the problem of representing the history of the learning process efficiently

January 23, 2015 17:58 ,

Doshi-Velez, F.; Pfau, D.; Wood, F.; Roy, N., Bayesian Nonparametric Methods for Partially-Observable Reinforcement Learning, Pattern Analysis and Machine Intelligence, IEEE Transactions on , vol.37, no.2, pp.394,407, Feb. 2015, DOI: 10.1109/TPAMI.2013.191

Making intelligent decisions from incomplete information is critical in many applications: for example, robots must choose actions based on imperfect sensors, and speech-based interfaces must infer a user\u2019s needs from noisy microphone inputs. What makes these tasks hard is that often we do not have a natural representation with which to model the domain and use for choosing actions; we must learn about the domain\u2019s properties while simultaneously performing the task. Learning a representation also involves trade-offs between modeling the data that we have seen previously and being able to make predictions about new data. This article explores learning representations of stochastic systems using Bayesian nonparametric statistics. Bayesian nonparametric methods allow the sophistication of a representation to scale gracefully with the complexity in the data. Our main contribution is a careful empirical evaluation of how representations learned using Bayesian nonparametric methods compare to other standard learning approaches, especially in support of planning and control. We show that the Bayesian aspects of the methods result in achieving state-of-the-art performance in decision making with relatively few samples, while the nonparametric aspects often result in fewer computations. These results hold across a variety of different techniques for choosing actions given a representation.

Posted in: Reinforcement learning theory , Tagged: ,

Reinforcement learning for discovering the parameters of the physical model of a system

December 5, 2014 17:08 ,

S.P. Nageshrao, G.A.D. Lopes, D. Jeltsema, R. Babuška, Passivity-based reinforcement learning control of a 2-DOF manipulator arm, Mechatronics, Volume 24, Issue 8, December 2014, Pages 1001-1007, ISSN 0957-4158, DOI: 10.1016/j.mechatronics.2014.10.005.

Passivity-based control (PBC) is commonly used for the stabilization of port-Hamiltonian (PH) systems. The PH framework is suitable for multi-domain systems, for example mechatronic devices or micro-electro-mechanical systems. Passivity-based control synthesis for PH systems involves solving partial differential equations, which can be cumbersome. Rather than explicitly solving these equations, in our approach the control law is parameterized and the unknown parameter vector is learned using an actor\u2013critic reinforcement learning algorithm. The key advantages of combining learning with PBC are: (i) the complexity of the control design procedure is reduced, (ii) prior knowledge about the system, given in the form of a PH model, speeds up the learning process, (iii) physical meaning can be attributed to the learned control law. In this paper we extended the learning-based PBC method to a regulation problem and present the experimental results for a two-degree-of-freedom manipulator. We show that the learning algorithm is capable of achieving feedback regulation in the presence of model uncertainties.

Posted in: Applications of reinforcement learning to control engineering, Control Engineering , Tagged: ,

A reinforcement learning controller to tune sub-controllers

December 5, 2014 17:02 ,

Kevin Van Vaerenbergh, Peter Vrancx, Yann-Michaël De Hauwere, Ann Nowé, Erik Hostens, Christophe Lauwerys, Tuning hydrostatic two-output drive-train controllers using reinforcement learning, Mechatronics, Volume 24, Issue 8, December 2014, Pages 975-985, ISSN 0957-4158. DOI: 10.1016/j.mechatronics.2014.07.005

When controlling a complex system consisting of several subsystems, a simple divide and conquer approach is to design a controller for each system separately. However, this does not necessarily result in a good overall control behavior. Especially when there are strong interactions between the subsystems, the selfish behavior of one controller might deteriorate the performance of the other subsystems. An alternative approach is to design a global controller for the entire mechatronic system. Such a design procedure might result in more optimal behavior, however it requires a lot more effort, especially when the interactions between the different subsystems cannot be modeled exactly or if the number of parameters is large.
In this paper we present a hybrid approach to this problem that overcomes the problems encountered when using several independent subsystems. Starting from such a system with individual subsystem controllers, we add a global layer which uses reinforcement learning to simultaneously tune the lower level controllers. While each subsystem still has its own individual controller, the reinforcement learning layer is used to tune these controllers in order to optimize global system behavior. This mitigates both the problem of subsystems behaving selfishly without the added complexity of designing a global controller for the entire system. Our approach is validated on a hydrostatic drive train.

Posted in: Applications of reinforcement learning to control engineering, Control Engineering , Tagged: ,

A new variant of Q-learning that alleviates its slow learning speed (with a brief review of reinforcement learning algorithms)

December 5, 2014 13:33 ,

J.C. van Rooijen, I. Grondman, R. Babuška, Learning rate free reinforcement learning for real-time motion control using a value-gradient based policy, Mechatronics, Volume 24, Issue 8, December 2014, Pages 966-974, ISSN 0957-4158. DOI: 10.1016/j.mechatronics.2014.05.007

Reinforcement learning (RL) is a framework that enables a controller to find an optimal control policy for a task in an unknown environment. Although RL has been successfully used to solve optimal control problems, learning is generally slow. The main causes are the inefficient use of information collected during interaction with the system and the inability to use prior knowledge on the system or the control task. In addition, the learning speed heavily depends on the learning rate parameter, which is difficult to tune.
In this paper, we present a sample-efficient, learning-rate-free version of the Value-Gradient Based Policy (VGBP) algorithm. The main difference between VGBP and other frequently used algorithms, such as Sarsa, is that in VGBP the learning agent has a direct access to the reward function, rather than just the immediate reward values. Furthermore, the agent learns a process model. This enables the algorithm to select control actions by optimizing over the right-hand side of the Bellman equation. We demonstrate the fast learning convergence in simulations and experiments with the underactuated pendulum swing-up task. In addition, we present experimental results for a more complex 2-DOF robotic manipulator.

Posted in: Applications of reinforcement learning to robots, Developmental robotics , Tagged: ,