- Showing results for:
- Reset all filters
Conference paperAlAttar A, Rouillard L, Kormushev P, 2019,
Autonomous air-hockey playing cobot using optimal control and vision-based Bayesian tracking, Towards Autonomous Robotic Systems, Publisher: Springer, ISSN: 0302-9743
This paper presents a novel autonomous air-hockey playing collaborative robot (cobot) that provides human-like gameplay against human opponents. Vision-based Bayesian tracking of the puck and striker are used in an Analytic Hierarchy Process (AHP)-based probabilistic tactical layer for high-speed perception. The tactical layer provides commands for an active control layer that controls the Cartesian position and yaw angle of a custom end effector. The active layer uses optimal control of the cobot’s posture inside the task nullspace. The kinematic redundancy is resolved using a weighted Moore-Penrose pseudo-inversion technique. Experiments with human players show high-speed human-like gameplay with potential applications in the growing field of entertainment robotics.
Journal articleMoriconi R, Kumar KSS, Deisenroth MP, 2020,
High-dimensional Bayesian optimization with projections using quantile Gaussian processes, Optimization Letters, ISSN: 1862-4472
Journal articleCreswell A, Bharath AA, 2019,
Denoising adversarial autoencoders, IEEE Transactions on Neural Networks and Learning Systems, Vol: 30, Pages: 968-984, ISSN: 2162-2388
Unsupervised learning is of growing interest because it unlocks the potential held in vast amounts of unlabeled data to learn useful representations for inference. Autoencoders, a form of generative model, may be trained by learning to reconstruct unlabeled input data from a latent representation space. More robust representations may be produced by an autoencoder if it learns to recover clean input samples from corrupted ones. Representations may be further improved by introducing regularization during training to shape the distribution of the encoded data in the latent space. We suggest denoising adversarial autoencoders (AAEs), which combine denoising and regularization, shaping the distribution of latent space using adversarial training. We introduce a novel analysis that shows how denoising may be incorporated into the training and sampling of AAEs. Experiments are performed to assess the contributions that denoising makes to the learning of representations for classification and sample synthesis. Our results suggest that autoencoders trained using a denoising criterion achieve higher classification performance and can synthesize samples that are more consistent with the input data than those trained without a corruption process.
Journal articleBertone G, Deisenroth MP, Kim JS, et al., 2019,
Accelerating the BSM interpretation of LHC data with machine learning, PHYSICS OF THE DARK UNIVERSE, Vol: 24, ISSN: 2212-6864
Journal articleKormushev P, Ugurlu B, Caldwell DG, et al., 2019,
Learning to exploit passive compliance for energy-efficient gait generation on a compliant humanoid, Autonomous Robots, Vol: 43, Pages: 79-95, ISSN: 0929-5593
Modern humanoid robots include not only active compliance but also passive compliance. Apart from improved safety and dependability, availability of passive elements, such as springs, opens up new possibilities for improving the energy efficiency. With this in mind, this paper addresses the challenging open problem of exploiting the passive compliance for the purpose of energy efficient humanoid walking. To this end, we develop a method comprising two parts: an optimization part that finds an optimal vertical center-of-mass trajectory, and a walking pattern generator part that uses this trajectory to produce a dynamically-balanced gait. For the optimization part, we propose a reinforcement learning approach that dynamically evolves the policy parametrization during the learning process. By gradually increasing the representational power of the policy parametrization, it manages to find better policies in a faster and computationally efficient way. For the walking generator part, we develop a variable-center-of-mass-height ZMP-based bipedal walking pattern generator. The method is tested in real-world experiments with the bipedal robot COMAN and achieves a significant 18% reduction in the electric energy consumption by learning to efficiently use the passive compliance of the robot.
Conference paperDutordoir V, Salimbeni HR, Hensman J, et al., 2018,
Gaussian process conditional density estimation, Advances in Neural Information Processing Systems, Publisher: Neural Information Processing Systems Conference
Conditional Density Estimation (CDE) models deal with estimating conditional distributions. The conditions imposed on the distribution are the inputs of the model. CDE is a challenging task as there is a fundamental trade-off between model complexity, representational capacity and overfitting. In this work, we propose to extend the model's input with latent variables and use Gaussian processes (GP) to map this augmented input onto samples from the conditional distribution. Our Bayesian approach allows for the modeling of small datasets, but we also provide the machinery for it to be applied to big data using stochastic variational inference. Our approach can be used to model densities even in sparse data regions, and allows for sharing learned structure between conditions. We illustrate the effectiveness and wide-reaching applicability of our model on a variety of real-world problems, such as spatio-temporal density estimation of taxi drop-offs, non-Gaussian noise modeling, and few-shot learning on omniglot images.
Conference paperWang K, Shah A, Kormushev P, 2018,
SLIDER: A Bipedal Robot with Knee-less Legs and Vertical Hip Sliding Motion, 21st International Conference on Climbing and Walking Robots and Support Technologies for Mobile Machines (CLAWAR 2018)
Conference paperPardo F, Levdik V, Kormushev P, 2018,
Q-map: A convolutional approach for goal-oriented reinforcement learning.
Goal-oriented learning has become a core concept in reinforcement learning(RL), extending the reward signal as a sole way to define tasks. However, asparameterizing value functions with goals increases the learning complexity,efficiently reusing past experience to update estimates towards several goalsat once becomes desirable but usually requires independent updates per goal.Considering that a significant number of RL environments can support spatialcoordinates as goals, such as on-screen location of the character in ATARI orSNES games, we propose a novel goal-oriented agent called Q-map that utilizesan autoencoder-like neural network to predict the minimum number of stepstowards each coordinate in a single forward pass. This architecture is similarto Horde with parameter sharing and allows the agent to discover correlationsbetween visual patterns and navigation. For example learning how to use aladder in a game could be transferred to other ladders later. We show how thisnetwork can be efficiently trained with a 3D variant of Q-learning to updatethe estimates towards all goals at once. While the Q-map agent could be usedfor a wide range of applications, we propose a novel exploration mechanism inplace of epsilon-greedy that relies on goal selection at a desired distancefollowed by several steps taken towards it, allowing long and coherentexploratory steps in the environment. We demonstrate the accuracy andgeneralization qualities of the Q-map agent on a grid-world environment andthen demonstrate the efficiency of the proposed exploration mechanism on thenotoriously difficult Montezuma's Revenge and Super Mario All-Stars games.
Conference paperSaputra RP, Kormushev P, 2018,
Casualty detection from 3D point cloud data for autonomous ground mobile rescue robots, SSRR 2018, Publisher: IEEE
One of the most important features of mobilerescue robots is the ability to autonomously detect casualties,i.e. human bodies, which are usually lying on the ground. Thispaper proposes a novel method for autonomously detectingcasualties lying on the ground using obtained 3D point-clouddata from an on-board sensor, such as an RGB-D camera ora 3D LIDAR, on a mobile rescue robot. In this method, theobtained 3D point-cloud data is projected onto the detectedground plane, i.e. floor, within the point cloud. Then, thisprojected point cloud is converted into a grid-map that isused afterwards as an input for the algorithm to detecthuman body shapes. The proposed method is evaluated byperforming detections of a human dummy, placed in differentrandom positions and orientations, using an on-board RGB-Dcamera on a mobile rescue robot called ResQbot. To evaluatethe robustness of the casualty detection method to differentcamera angles, the orientation of the camera is set to differentangles. The experimental results show that using the point-clouddata from the on-board RGB-D camera, the proposed methodsuccessfully detects the casualty in all tested body positions andorientations relative to the on-board camera, as well as in alltested camera angles.
Conference paperSalimbeni HR, Cheng C-A, Boots B, et al., 2018,
Orthogonally decoupled variational Gaussian processes, Advances in Neural Information Processing Systems (NIPS) 2018, Publisher: Massachusetts Institute of Technology Press, ISSN: 1049-5258
Gaussian processes (GPs) provide a powerful non-parametric framework for rea-soning over functions. Despite appealing theory, its superlinear computational andmemory complexities have presented a long-standing challenge. State-of-the-artsparse variational inference methods trade modeling accuracy against complexity.However, the complexities of these methods still scale superlinearly in the numberof basis functions, implying that that sparse GP methods are able to learn fromlarge datasets only when a small model is used. Recently, a decoupled approachwas proposed that removes the unnecessary coupling between the complexitiesof modeling the mean and the covariance functions of a GP. It achieves a linearcomplexity in the number of mean parameters, so an expressive posterior meanfunction can be modeled. While promising, this approach suffers from optimizationdifficulties due to ill-conditioning and non-convexity. In this work, we propose analternative decoupled parametrization. It adopts an orthogonal basis in the meanfunction to model the residues that cannot be learned by the standard coupled ap-proach. Therefore, our method extends, rather than replaces, the coupled approachto achieve strictly better performance. This construction admits a straightforwardnatural gradient update rule, so the structure of the information manifold that islost during decoupling can be leveraged to speed up learning. Empirically, ouralgorithm demonstrates significantly faster convergence in multiple experiments.
This data is extracted from the Web of Science and reproduced under a licence from Thomson Reuters. You may not copy or re-distribute this data in whole or in part without the written consent of the Science business of Thomson Reuters.