Header logo is


2017


Thumb xl larsnips
The Numerics of GANs

Mescheder, L., Nowozin, S., Geiger, A.

In Proceedings from the conference "Neural Information Processing Systems 2017., (Editors: Guyon I. and Luxburg U.v. and Bengio S. and Wallach H. and Fergus R. and Vishwanathan S. and Garnett R.), Curran Associates, Inc., Advances in Neural Information Processing Systems 30 (NIPS), December 2017 (inproceedings)

Abstract
In this paper, we analyze the numerics of common algorithms for training Generative Adversarial Networks (GANs). Using the formalism of smooth two-player games we analyze the associated gradient vector field of GAN training objectives. Our findings suggest that the convergence of current algorithms suffers due to two factors: i) presence of eigenvalues of the Jacobian of the gradient vector field with zero real-part, and ii) eigenvalues with big imaginary part. Using these findings, we design a new algorithm that overcomes some of these limitations and has better convergence properties. Experimentally, we demonstrate its superiority on training common GAN architectures and show convergence on GAN architectures that are known to be notoriously hard to train.

avg

pdf Project Page [BibTex]

2017


pdf Project Page [BibTex]


Thumb xl amd intentiongan
Multi-Modal Imitation Learning from Unstructured Demonstrations using Generative Adversarial Nets

Hausman, K., Chebotar, Y., Schaal, S., Sukhatme, G., Lim, J.

In Proceedings from the conference "Neural Information Processing Systems 2017., (Editors: Guyon I. and Luxburg U.v. and Bengio S. and Wallach H. and Fergus R. and Vishwanathan S. and Garnett R.), Curran Associates, Inc., Advances in Neural Information Processing Systems 30 (NIPS), December 2017 (inproceedings)

am

pdf video [BibTex]

pdf video [BibTex]


Thumb xl fig toyex lqr1kernel 1
On the Design of LQR Kernels for Efficient Controller Learning

Marco, A., Hennig, P., Schaal, S., Trimpe, S.

Proceedings of the 56th IEEE Annual Conference on Decision and Control (CDC), pages: 5193-5200, IEEE, IEEE Conference on Decision and Control, December 2017 (conference)

Abstract
Finding optimal feedback controllers for nonlinear dynamic systems from data is hard. Recently, Bayesian optimization (BO) has been proposed as a powerful framework for direct controller tuning from experimental trials. For selecting the next query point and finding the global optimum, BO relies on a probabilistic description of the latent objective function, typically a Gaussian process (GP). As is shown herein, GPs with a common kernel choice can, however, lead to poor learning outcomes on standard quadratic control problems. For a first-order system, we construct two kernels that specifically leverage the structure of the well-known Linear Quadratic Regulator (LQR), yet retain the flexibility of Bayesian nonparametric learning. Simulations of uncertain linear and nonlinear systems demonstrate that the LQR kernels yield superior learning performance.

am ics pn

arXiv PDF On the Design of LQR Kernels for Efficient Controller Learning - CDC presentation DOI Project Page [BibTex]

arXiv PDF On the Design of LQR Kernels for Efficient Controller Learning - CDC presentation DOI Project Page [BibTex]


Thumb xl teaser
Optimizing Long-term Predictions for Model-based Policy Search

Doerr, A., Daniel, C., Nguyen-Tuong, D., Marco, A., Schaal, S., Toussaint, M., Trimpe, S.

Proceedings of 1st Annual Conference on Robot Learning (CoRL), 78, pages: 227-238, (Editors: Sergey Levine and Vincent Vanhoucke and Ken Goldberg), 1st Annual Conference on Robot Learning, November 2017 (conference)

Abstract
We propose a novel long-term optimization criterion to improve the robustness of model-based reinforcement learning in real-world scenarios. Learning a dynamics model to derive a solution promises much greater data-efficiency and reusability compared to model-free alternatives. In practice, however, modelbased RL suffers from various imperfections such as noisy input and output data, delays and unmeasured (latent) states. To achieve higher resilience against such effects, we propose to optimize a generative long-term prediction model directly with respect to the likelihood of observed trajectories as opposed to the common approach of optimizing a dynamics model for one-step-ahead predictions. We evaluate the proposed method on several artificial and real-world benchmark problems and compare it to PILCO, a model-based RL framework, in experiments on a manipulation robot. The results show that the proposed method is competitive compared to state-of-the-art model learning methods. In contrast to these more involved models, our model can directly be employed for policy search and outperforms a baseline method in the robot experiment.

am ics

PDF Project Page [BibTex]

PDF Project Page [BibTex]


Thumb xl teaser iccv2017
Bounding Boxes, Segmentations and Object Coordinates: How Important is Recognition for 3D Scene Flow Estimation in Autonomous Driving Scenarios?

Behl, A., Jafari, O. H., Mustikovela, S. K., Alhaija, H. A., Rother, C., Geiger, A.

In Proceedings IEEE International Conference on Computer Vision (ICCV), IEEE, Piscataway, NJ, USA, IEEE International Conference on Computer Vision (ICCV), October 2017 (inproceedings)

Abstract
Existing methods for 3D scene flow estimation often fail in the presence of large displacement or local ambiguities, e.g., at texture-less or reflective surfaces. However, these challenges are omnipresent in dynamic road scenes, which is the focus of this work. Our main contribution is to overcome these 3D motion estimation problems by exploiting recognition. In particular, we investigate the importance of recognition granularity, from coarse 2D bounding box estimates over 2D instance segmentations to fine-grained 3D object part predictions. We compute these cues using CNNs trained on a newly annotated dataset of stereo images and integrate them into a CRF-based model for robust 3D scene flow estimation - an approach we term Instance Scene Flow. We analyze the importance of each recognition cue in an ablation study and observe that the instance segmentation cue is by far strongest, in our setting. We demonstrate the effectiveness of our method on the challenging KITTI 2015 scene flow benchmark where we achieve state-of-the-art performance at the time of submission.

avg

pdf suppmat Poster Project Page [BibTex]

pdf suppmat Poster Project Page [BibTex]


Thumb xl jonas teaser
Sparsity Invariant CNNs

Uhrig, J., Schneider, N., Schneider, L., Franke, U., Brox, T., Geiger, A.

International Conference on 3D Vision (3DV) 2017, International Conference on 3D Vision (3DV), October 2017 (conference)

Abstract
In this paper, we consider convolutional neural networks operating on sparse inputs with an application to depth upsampling from sparse laser scan data. First, we show that traditional convolutional networks perform poorly when applied to sparse data even when the location of missing data is provided to the network. To overcome this problem, we propose a simple yet effective sparse convolution layer which explicitly considers the location of missing data during the convolution operation. We demonstrate the benefits of the proposed network architecture in synthetic and real experiments \wrt various baseline approaches. Compared to dense baselines, the proposed sparse convolution network generalizes well to novel datasets and is invariant to the level of sparsity in the data. For our evaluation, we derive a novel dataset from the KITTI benchmark, comprising 93k depth annotated RGB images. Our dataset allows for training and evaluating depth upsampling and depth prediction techniques in challenging real-world settings.

avg

pdf suppmat Project Page Project Page [BibTex]

pdf suppmat Project Page Project Page [BibTex]


Thumb xl gernot teaser
OctNetFusion: Learning Depth Fusion from Data

Riegler, G., Ulusoy, A. O., Bischof, H., Geiger, A.

International Conference on 3D Vision (3DV) 2017, International Conference on 3D Vision (3DV), October 2017 (conference)

Abstract
In this paper, we present a learning based approach to depth fusion, i.e., dense 3D reconstruction from multiple depth images. The most common approach to depth fusion is based on averaging truncated signed distance functions, which was originally proposed by Curless and Levoy in 1996. While this method is simple and provides great results, it is not able to reconstruct (partially) occluded surfaces and requires a large number frames to filter out sensor noise and outliers. Motivated by the availability of large 3D model repositories and recent advances in deep learning, we present a novel 3D CNN architecture that learns to predict an implicit surface representation from the input depth maps. Our learning based method significantly outperforms the traditional volumetric fusion approach in terms of noise reduction and outlier suppression. By learning the structure of real world 3D objects and scenes, our approach is further able to reconstruct occluded regions and to fill in gaps in the reconstruction. We demonstrate that our learning based approach outperforms both vanilla TSDF fusion as well as TV-L1 fusion on the task of volumetric fusion. Further, we demonstrate state-of-the-art 3D shape completion results.

avg

pdf Video 1 Video 2 Project Page Project Page [BibTex]

pdf Video 1 Video 2 Project Page Project Page [BibTex]


Thumb xl andreas teaser
Direct Visual Odometry for a Fisheye-Stereo Camera

Liu, P., Heng, L., Sattler, T., Geiger, A., Pollefeys, M.

In Proceedings IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), IEEE, Piscataway, NJ, USA, IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), September 2017 (inproceedings)

Abstract
We present a direct visual odometry algorithm for a fisheye-stereo camera. Our algorithm performs simultaneous camera motion estimation and semi-dense reconstruction. The pipeline consists of two threads: a tracking thread and a mapping thread. In the tracking thread, we estimate the camera pose via semi-dense direct image alignment. To have a wider field of view (FoV) which is important for robotic perception, we use fisheye images directly without converting them to conventional pinhole images which come with a limited FoV. To address the epipolar curve problem, plane-sweeping stereo is used for stereo matching and depth initialization. Multiple depth hypotheses are tracked for selected pixels to better capture the uncertainty characteristics of stereo matching. Temporal motion stereo is then used to refine the depth and remove false positive depth hypotheses. Our implementation runs at an average of 20 Hz on a low-end PC. We run experiments in outdoor environments to validate our algorithm, and discuss the experimental results. We experimentally show that we are able to estimate 6D poses with low drift, and at the same time, do semi-dense 3D reconstruction with high accuracy.

avg

pdf Project Page [BibTex]

pdf Project Page [BibTex]


no image
A New Data Source for Inverse Dynamics Learning

Kappler, D., Meier, F., Ratliff, N., Schaal, S.

In Proceedings IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), IEEE, Piscataway, NJ, USA, IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), September 2017 (inproceedings)

am

[BibTex]

[BibTex]


no image
Bayesian Regression for Artifact Correction in Electroencephalography

Fiebig, K., Jayaram, V., Hesse, T., Blank, A., Peters, J., Grosse-Wentrup, M.

Proceedings of the 7th Graz Brain-Computer Interface Conference 2017 - From Vision to Reality, pages: 131-136, (Editors: Müller-Putz G.R., Steyrl D., Wriessnegger S. C., Scherer R.), Graz University of Technology, Austria, Graz Brain-Computer Interface Conference, September 2017 (conference)

am ei

DOI [BibTex]

DOI [BibTex]


no image
Investigating Music Imagery as a Cognitive Paradigm for Low-Cost Brain-Computer Interfaces

Grossberger, L., Hohmann, M. R., Peters, J., Grosse-Wentrup, M.

Proceedings of the 7th Graz Brain-Computer Interface Conference 2017 - From Vision to Reality, pages: 160-164, (Editors: Müller-Putz G.R., Steyrl D., Wriessnegger S. C., Scherer R.), Graz University of Technology, Austria, Graz Brain-Computer Interface Conference, September 2017 (conference)

am ei

DOI [BibTex]

DOI [BibTex]


Thumb xl screen shot 2017 08 01 at 15.41.10
On the relevance of grasp metrics for predicting grasp success

Rubert, C., Kappler, D., Morales, A., Schaal, S., Bohg, J.

In Proceedings of the IEEE/RSJ International Conference of Intelligent Robots and Systems, September 2017 (inproceedings) Accepted

Abstract
We aim to reliably predict whether a grasp on a known object is successful before it is executed in the real world. There is an entire suite of grasp metrics that has already been developed which rely on precisely known contact points between object and hand. However, it remains unclear whether and how they may be combined into a general purpose grasp stability predictor. In this paper, we analyze these questions by leveraging a large scale database of simulated grasps on a wide variety of objects. For each grasp, we compute the value of seven metrics. Each grasp is annotated by human subjects with ground truth stability labels. Given this data set, we train several classification methods to find out whether there is some underlying, non-trivial structure in the data that is difficult to model manually but can be learned. Quantitative and qualitative results show the complexity of the prediction problem. We found that a good prediction performance critically depends on using a combination of metrics as input features. Furthermore, non-parametric and non-linear classifiers best capture the structure in the data.

am

Project Page [BibTex]

Project Page [BibTex]


Thumb xl hassan paper teasere
Augmented Reality Meets Deep Learning for Car Instance Segmentation in Urban Scenes

Alhaija, H. A., Mustikovela, S. K., Mescheder, L., Geiger, A., Rother, C.

In Proceedings of the British Machine Vision Conference 2017, Proceedings of the British Machine Vision Conference, September 2017 (inproceedings)

Abstract
The success of deep learning in computer vision is based on the availability of large annotated datasets. To lower the need for hand labeled images, virtually rendered 3D worlds have recently gained popularity. Unfortunately, creating realistic 3D content is challenging on its own and requires significant human effort. In this work, we propose an alternative paradigm which combines real and synthetic data for learning semantic instance segmentation models. Exploiting the fact that not all aspects of the scene are equally important for this task, we propose to augment real-world imagery with virtual objects of the target category. Capturing real-world images at large scale is easy and cheap, and directly provides real background appearances without the need for creating complex 3D models of the environment. We present an efficient procedure to augment these images with virtual objects. This allows us to create realistic composite images which exhibit both realistic background appearance as well as a large number of complex object arrangements. In contrast to modeling complete 3D environments, our data augmentation approach requires only a few user interactions in combination with 3D shapes of the target object category. We demonstrate the utility of the proposed approach for training a state-of-the-art high-capacity deep model for semantic instance segmentation. In particular, we consider the task of segmenting car instances on the KITTI dataset which we have annotated with pixel-accurate ground truth. Our experiments demonstrate that models trained on augmented imagery generalize better than those trained on synthetic data or models trained on limited amounts of annotated real data.

avg

pdf Project Page [BibTex]

pdf Project Page [BibTex]


no image
Local Bayesian Optimization of Motor Skills

Akrour, R., Sorokin, D., Peters, J., Neumann, G.

Proceedings of the 34th International Conference on Machine Learning, 70, pages: 41-50, Proceedings of Machine Learning Research, (Editors: Doina Precup, Yee Whye Teh), PMLR, International Conference on Machine Learning (ICML), August 2017 (conference)

am ei

link (url) Project Page [BibTex]

link (url) Project Page [BibTex]


Thumb xl img01
Adversarial Variational Bayes: Unifying Variational Autoencoders and Generative Adversarial Networks

Mescheder, L., Nowozin, S., Geiger, A.

In Proceedings of the 34th International Conference on Machine Learning, 70, Proceedings of Machine Learning Research, (Editors: Doina Precup, Yee Whye Teh), PMLR, International Conference on Machine Learning (ICML), August 2017 (inproceedings)

Abstract
Variational Autoencoders (VAEs) are expressive latent variable models that can be used to learn complex probability distributions from training data. However, the quality of the resulting model crucially relies on the expressiveness of the inference model. We introduce Adversarial Variational Bayes (AVB), a technique for training Variational Autoencoders with arbitrarily expressive inference models. We achieve this by introducing an auxiliary discriminative network that allows to rephrase the maximum-likelihood-problem as a two-player game, hence establishing a principled connection between VAEs and Generative Adversarial Networks (GANs). We show that in the nonparametric limit our method yields an exact maximum-likelihood assignment for the parameters of the generative model, as well as the exact posterior distribution over the latent variables given an observation. Contrary to competing approaches which combine VAEs with GANs, our approach has a clear theoretical justification, retains most advantages of standard Variational Autoencoders and is easy to implement.

avg

pdf suppmat Project Page arxiv-version Project Page [BibTex]

pdf suppmat Project Page arxiv-version Project Page [BibTex]


Thumb xl pilqr cover
Combining Model-Based and Model-Free Updates for Trajectory-Centric Reinforcement Learning

Chebotar, Y., Hausman, K., Zhang, M., Sukhatme, G., Schaal, S., Levine, S.

Proceedings of the 34th International Conference on Machine Learning, 70, Proceedings of Machine Learning Research, (Editors: Doina Precup, Yee Whye Teh), PMLR, International Conference on Machine Learning (ICML), August 2017 (conference)

am

pdf video [BibTex]

pdf video [BibTex]


Thumb xl joel slow flow crop
Slow Flow: Exploiting High-Speed Cameras for Accurate and Diverse Optical Flow Reference Data

Janai, J., Güney, F., Wulff, J., Black, M., Geiger, A.

In Proceedings IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2017, pages: 1406-1416, IEEE, Piscataway, NJ, USA, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), July 2017 (inproceedings)

Abstract
Existing optical flow datasets are limited in size and variability due to the difficulty of capturing dense ground truth. In this paper, we tackle this problem by tracking pixels through densely sampled space-time volumes recorded with a high-speed video camera. Our model exploits the linearity of small motions and reasons about occlusions from multiple frames. Using our technique, we are able to establish accurate reference flow fields outside the laboratory in natural environments. Besides, we show how our predictions can be used to augment the input images with realistic motion blur. We demonstrate the quality of the produced flow fields on synthetic and real-world datasets. Finally, we collect a novel challenging optical flow dataset by applying our technique on data from a high-speed camera and analyze the performance of the state-of-the-art in optical flow under various levels of motion blur.

avg ps

pdf suppmat Project page Video DOI Project Page [BibTex]

pdf suppmat Project page Video DOI Project Page [BibTex]


Thumb xl img03
OctNet: Learning Deep 3D Representations at High Resolutions

Riegler, G., Ulusoy, O., Geiger, A.

In Proceedings IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2017, IEEE, Piscataway, NJ, USA, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), July 2017 (inproceedings)

Abstract
We present OctNet, a representation for deep learning with sparse 3D data. In contrast to existing models, our representation enables 3D convolutional networks which are both deep and high resolution. Towards this goal, we exploit the sparsity in the input data to hierarchically partition the space using a set of unbalanced octrees where each leaf node stores a pooled feature representation. This allows to focus memory allocation and computation to the relevant dense regions and enables deeper networks without compromising resolution. We demonstrate the utility of our OctNet representation by analyzing the impact of resolution on several 3D tasks including 3D object classification, orientation estimation and point cloud labeling.

avg ps

pdf suppmat Project Page Video Project Page [BibTex]

pdf suppmat Project Page Video Project Page [BibTex]


Thumb xl schoeps2017cvpr
A Multi-View Stereo Benchmark with High-Resolution Images and Multi-Camera Videos

Schöps, T., Schönberger, J. L., Galliani, S., Sattler, T., Schindler, K., Pollefeys, M., Geiger, A.

In Proceedings IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2017, IEEE, Piscataway, NJ, USA, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), July 2017 (inproceedings)

Abstract
Motivated by the limitations of existing multi-view stereo benchmarks, we present a novel dataset for this task. Towards this goal, we recorded a variety of indoor and outdoor scenes using a high-precision laser scanner and captured both high-resolution DSLR imagery as well as synchronized low-resolution stereo videos with varying fields-of-view. To align the images with the laser scans, we propose a robust technique which minimizes photometric errors conditioned on the geometry. In contrast to previous datasets, our benchmark provides novel challenges and covers a diverse set of viewpoints and scene types, ranging from natural scenes to man-made indoor and outdoor environments. Furthermore, we provide data at significantly higher temporal and spatial resolution. Our benchmark is the first to cover the important use case of hand-held mobile devices while also providing high-resolution DSLR camera images. We make our datasets and an online evaluation server available at http://www.eth3d.net.

avg

pdf suppmat Project Page Project Page [BibTex]

pdf suppmat Project Page Project Page [BibTex]


Thumb xl camposeco2017cvpr
Toroidal Constraints for Two Point Localization Under High Outlier Ratios

Camposeco, F., Sattler, T., Cohen, A., Geiger, A., Pollefeys, M.

In Proceedings IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2017, IEEE, Piscataway, NJ, USA, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), July 2017 (inproceedings)

Abstract
Localizing a query image against a 3D model at large scale is a hard problem, since 2D-3D matches become more and more ambiguous as the model size increases. This creates a need for pose estimation strategies that can handle very low inlier ratios. In this paper, we draw new insights on the geometric information available from the 2D-3D matching process. As modern descriptors are not invariant against large variations in viewpoint, we are able to find the rays in space used to triangulate a given point that are closest to a query descriptor. It is well known that two correspondences constrain the camera to lie on the surface of a torus. Adding the knowledge of direction of triangulation, we are able to approximate the position of the camera from \emphtwo matches alone. We derive a geometric solver that can compute this position in under 1 microsecond. Using this solver, we propose a simple yet powerful outlier filter which scales quadratically in the number of matches. We validate the accuracy of our solver and demonstrate the usefulness of our method in real world settings.

avg

pdf suppmat Project Page Project Page [BibTex]

pdf suppmat Project Page pdf Project Page [BibTex]


Thumb xl cvpr2017 landpsace
Semantic Multi-view Stereo: Jointly Estimating Objects and Voxels

Ulusoy, A. O., Black, M. J., Geiger, A.

In Proceedings IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2017, IEEE, Piscataway, NJ, USA, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), July 2017 (inproceedings)

Abstract
Dense 3D reconstruction from RGB images is a highly ill-posed problem due to occlusions, textureless or reflective surfaces, as well as other challenges. We propose object-level shape priors to address these ambiguities. Towards this goal, we formulate a probabilistic model that integrates multi-view image evidence with 3D shape information from multiple objects. Inference in this model yields a dense 3D reconstruction of the scene as well as the existence and precise 3D pose of the objects in it. Our approach is able to recover fine details not captured in the input shapes while defaulting to the input models in occluded regions where image evidence is weak. Due to its probabilistic nature, the approach is able to cope with the approximate geometry of the 3D models as well as input shapes that are not present in the scene. We evaluate the approach quantitatively on several challenging indoor and outdoor datasets.

avg ps

YouTube pdf suppmat Project Page [BibTex]

YouTube pdf suppmat Project Page [BibTex]


Thumb xl apollo system2 croped
Model-Based Policy Search for Automatic Tuning of Multivariate PID Controllers

Doerr, A., Nguyen-Tuong, D., Marco, A., Schaal, S., Trimpe, S.

In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), pages: 5295-5301, IEEE, Piscataway, NJ, USA, IEEE International Conference on Robotics and Automation (ICRA), May 2017 (inproceedings)

am ics

PDF arXiv DOI Project Page [BibTex]

PDF arXiv DOI Project Page [BibTex]


Thumb xl learning ct block diagram v2
Learning Feedback Terms for Reactive Planning and Control

Rai, A., Sutanto, G., Schaal, S., Meier, F.

Proceedings 2017 IEEE International Conference on Robotics and Automation (ICRA), IEEE, Piscataway, NJ, USA, IEEE International Conference on Robotics and Automation (ICRA), May 2017 (conference)

am

pdf video [BibTex]

pdf video [BibTex]


Thumb xl this one
Virtual vs. Real: Trading Off Simulations and Physical Experiments in Reinforcement Learning with Bayesian Optimization

Marco, A., Berkenkamp, F., Hennig, P., Schoellig, A. P., Krause, A., Schaal, S., Trimpe, S.

In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), pages: 1557-1563, IEEE, Piscataway, NJ, USA, IEEE International Conference on Robotics and Automation (ICRA), May 2017 (inproceedings)

am ics pn

PDF arXiv ICRA 2017 Spotlight presentation Virtual vs. Real - Video explanation DOI Project Page [BibTex]

PDF arXiv ICRA 2017 Spotlight presentation Virtual vs. Real - Video explanation DOI Project Page [BibTex]

2006


no image
Learning operational space control

Peters, J., Schaal, S.

In Robotics: Science and Systems II (RSS 2006), pages: 255-262, (Editors: Gaurav S. Sukhatme and Stefan Schaal and Wolfram Burgard and Dieter Fox), Cambridge, MA: MIT Press, RSS , 2006, clmc (inproceedings)

Abstract
While operational space control is of essential importance for robotics and well-understood from an analytical point of view, it can be prohibitively hard to achieve accurate control in face of modeling errors, which are inevitable in complex robots, e.g., humanoid robots. In such cases, learning control methods can offer an interesting alternative to analytical control algorithms. However, the resulting learning problem is ill-defined as it requires to learn an inverse mapping of a usually redundant system, which is well known to suffer from the property of non-covexity of the solution space, i.e., the learning system could generate motor commands that try to steer the robot into physically impossible configurations. A first important insight for this paper is that, nevertheless, a physically correct solution to the inverse problem does exits when learning of the inverse map is performed in a suitable piecewise linear way. The second crucial component for our work is based on a recent insight that many operational space controllers can be understood in terms of a constraint optimal control problem. The cost function associated with this optimal control problem allows us to formulate a learning algorithm that automatically synthesizes a globally consistent desired resolution of redundancy while learning the operational space controller. From the view of machine learning, the learning problem corresponds to a reinforcement learning problem that maximizes an immediate reward and that employs an expectation-maximization policy search algorithm. Evaluations on a three degrees of freedom robot arm illustrate the feasability of our suggested approach.

am ei

link (url) [BibTex]

2006


link (url) [BibTex]


no image
Reinforcement Learning for Parameterized Motor Primitives

Peters, J., Schaal, S.

In Proceedings of the 2006 International Joint Conference on Neural Networks, pages: 73-80, IJCNN, 2006, clmc (inproceedings)

Abstract
One of the major challenges in both action generation for robotics and in the understanding of human motor control is to learn the "building blocks of movement generation", called motor primitives. Motor primitives, as used in this paper, are parameterized control policies such as splines or nonlinear differential equations with desired attractor properties. While a lot of progress has been made in teaching parameterized motor primitives using supervised or imitation learning, the self-improvement by interaction of the system with the environment remains a challenging problem. In this paper, we evaluate different reinforcement learning approaches for improving the performance of parameterized motor primitives. For pursuing this goal, we highlight the difficulties with current reinforcement learning methods, and outline both established and novel algorithms for the gradient-based improvement of parameterized policies. We compare these algorithms in the context of motor primitive learning, and show that our most modern algorithm, the Episodic Natural Actor-Critic outperforms previous algorithms by at least an order of magnitude. We demonstrate the efficiency of this reinforcement learning method in the application of learning to hit a baseball with an anthropomorphic robot arm.

am ei

link (url) DOI [BibTex]

link (url) DOI [BibTex]

2003


no image
Dynamic movement primitives - A framework for motor control in humans and humanoid robots

Schaal, S.

In The International Symposium on Adaptive Motion of Animals and Machines, Kyoto, Japan, March 4-8, 2003, March 2003, clmc (inproceedings)

Abstract
Sensory-motor integration is one of the key issues in robotics. In this paper, we propose an approach to rhythmic arm movement control that is synchronized with an external signal based on exploiting a simple neural oscillator network. Trajectory generation by the neural oscillator is a biologically inspired method that can allow us to generate a smooth and continuous trajectory. The parameter tuning of the oscillators is used to generate a synchronized movement with wide intervals. We adopted the method for the drumming task as an example task. By using this method, the robot can realize synchronized drumming with wide drumming intervals in real time. The paper also shows the experimental results of drumming by a humanoid robot.

am

link (url) [BibTex]

2003


link (url) [BibTex]


no image
Bayesian backfitting

D’Souza, A., Vijayakumar, S., Schaal, S.

In Proceedings of the 10th Joint Symposium on Neural Computation (JSNC 2003), Irvine, CA, May 2003, 2003, clmc (inproceedings)

Abstract
We present an algorithm aimed at addressing both computational and analytical intractability of Bayesian regression models which operate in very high-dimensional, usually underconstrained spaces. Several domains of research frequently provide such datasets, including chemometrics [2], and human movement analysis [1]. The literature in nonparametric statistics provides interesting solutions such as Backfitting [3] and Partial Least Squares [4], which are extremely robust and efficient, yet lack a probabilistic interpretation that could place them in the context of current research in statistical learning algorithms that emphasize the estimation of confidence, posterior distributions, and model complexity. In order to achieve numerical robustness and low computational cost, we first derive a novel Bayesian interpretation of Backfitting (BB) as a computationally efficient regression algorithm. BBÕs learning complexity scales linearly with the input dimensionality by decoupling inference among individual input dimensions. We embed BB in an efficient, locally variational model selection mechanism that automatically grows the number of backfitting experts in a mixture-of-experts regression model. We demonstrate the effectiveness of the algorithm in performing principled regularization of model complexity when fitting nonlinear manifolds while avoiding the numerical hazards associated with highly underconstrained problems. We also note that this algorithm appears applicable in various areas of neural computation, e.g., in abstract models of computational neuroscience, or implementations of statistical learning on artificial systems.

am

link (url) [BibTex]

link (url) [BibTex]


no image
Reinforcement learning for humanoid robotics

Peters, J., Vijayakumar, S., Schaal, S.

In IEEE-RAS International Conference on Humanoid Robots (Humanoids2003), Karlsruhe, Germany, Sept.29-30, 2003, clmc (inproceedings)

Abstract
Reinforcement learning offers one of the most general framework to take traditional robotics towards true autonomy and versatility. However, applying reinforcement learning to high dimensional movement systems like humanoid robots remains an unsolved problem. In this paper, we discuss different approaches of reinforcement learning in terms of their applicability in humanoid robotics. Methods can be coarsely classified into three different categories, i.e., greedy methods, `vanilla' policy gradient methods, and natural gradient methods. We discuss that greedy methods are not likely to scale into the domain humanoid robotics as they are problematic when used with function approximation. `Vanilla' policy gradient methods on the other hand have been successfully applied on real-world robots including at least one humanoid robot. We demonstrate that these methods can be significantly improved using the natural policy gradient instead of the regular policy gradient. A derivation of the natural policy gradient is provided, proving that the average policy gradient of Kakade (2002) is indeed the true natural gradient. A general algorithm for estimating the natural gradient, the Natural Actor-Critic algorithm, is introduced. This algorithm converges to the nearest local minimum of the cost function with respect to the Fisher information metric under suitable conditions. The algorithm outperforms non-natural policy gradients by far in a cart-pole balancing evaluation, and for learning nonlinear dynamic motor primitives for humanoid robot control. It offers a promising route for the development of reinforcement learning for truly high dimensionally continuous state-action systems.

am

link (url) [BibTex]

link (url) [BibTex]


no image
Discovering imitation strategies through categorization of multi-cimensional data

Billard, A., Epars, Y., Schaal, S., Cheng, G.

In IEEE International Conference on Intelligent Robots and Systems (IROS 2003), Las Vegas, NV, Oct. 27-31, 2003, clmc (inproceedings)

Abstract
An essential problem of imitation is that of determining Ówhat to imitateÓ, i.e. to determine which of the many features of the demonstration are relevant to the task and which should be reproduced. The strategy followed by the imitator can be modeled as a hierarchical optimization system, which minimizes the discrepancy between two multidimensional datasets. We consider imitation of a manipulation task. To classify across manipulation strategies, we apply a probabilistic analysis to data in Cartesian and joint spaces. We determine a general metric that optimizes the policy of task reproduction, following strategy determination. The model successfully discovers strategies in six different manipulation tasks and controls task reproduction by a full body humanoid robot. or the complete path followed by the demonstrator. We follow a similar taxonomy and apply it to the learning and reproduction of a manipulation task by a humanoid robot. We take the perspective that the features of the movements to imitate are those that appear most frequently, i.e. the invariants in time. The model builds upon previous work [3], [4] and is composed of a hierarchical time delay neural network that extracts invariant features from a manipulation task performed by a human demonstrator. The system analyzes the Carthesian trajectories of the objects and the joint

am

link (url) [BibTex]

link (url) [BibTex]


no image
Scaling reinforcement learning paradigms for motor learning

Peters, J., Vijayakumar, S., Schaal, S.

In Proceedings of the 10th Joint Symposium on Neural Computation (JSNC 2003), Irvine, CA, May 2003, 2003, clmc (inproceedings)

Abstract
Reinforcement learning offers a general framework to explain reward related learning in artificial and biological motor control. However, current reinforcement learning methods rarely scale to high dimensional movement systems and mainly operate in discrete, low dimensional domains like game-playing, artificial toy problems, etc. This drawback makes them unsuitable for application to human or bio-mimetic motor control. In this poster, we look at promising approaches that can potentially scale and suggest a novel formulation of the actor-critic algorithm which takes steps towards alleviating the current shortcomings. We argue that methods based on greedy policies are not likely to scale into high-dimensional domains as they are problematic when used with function approximation Ð a must when dealing with continuous domains. We adopt the path of direct policy gradient based policy improvements since they avoid the problems of unstabilizing dynamics encountered in traditional value iteration based updates. While regular policy gradient methods have demonstrated promising results in the domain of humanoid notor control, we demonstrate that these methods can be significantly improved using the natural policy gradient instead of the regular policy gradient. Based on this, it is proved that KakadeÕs Ôaverage natural policy gradientÕ is indeed the true natural gradient. A general algorithm for estimating the natural gradient, the Natural Actor-Critic algorithm, is introduced. This algorithm converges with probability one to the nearest local minimum in Riemannian space of the cost function. The algorithm outperforms nonnatural policy gradients by far in a cart-pole balancing evaluation, and offers a promising route for the development of reinforcement learning for truly high-dimensionally continuous state-action systems.

am

link (url) [BibTex]

link (url) [BibTex]


no image
Learning attractor landscapes for learning motor primitives

Ijspeert, A., Nakanishi, J., Schaal, S.

In Advances in Neural Information Processing Systems 15, pages: 1547-1554, (Editors: Becker, S.;Thrun, S.;Obermayer, K.), Cambridge, MA: MIT Press, 2003, clmc (inproceedings)

Abstract
If globally high dimensional data has locally only low dimensional distributions, it is advantageous to perform a local dimensionality reduction before further processing the data. In this paper we examine several techniques for local dimensionality reduction in the context of locally weighted linear regression. As possible candidates, we derive local versions of factor analysis regression, principle component regression, principle component regression on joint distributions, and partial least squares regression. After outlining the statistical bases of these methods, we perform Monte Carlo simulations to evaluate their robustness with respect to violations of their statistical assumptions. One surprising outcome is that locally weighted partial least squares regression offers the best average results, thus outperforming even factor analysis, the theoretically most appealing of our candidate techniques.Ê

am

link (url) [BibTex]

link (url) [BibTex]


no image
Learning from demonstration and adaptation of biped locomotion with dynamical movement primitives

Nakanishi, J., Morimoto, J., Endo, G., Schaal, S., Kawato, M.

In Workshop on Robot Learning by Demonstration, IEEE International Conference on Intelligent Robots and Systems (IROS 2003), Las Vegas, NV, Oct. 27-31, 2003, clmc (inproceedings)

Abstract
In this paper, we report on our research for learning biped locomotion from human demonstration. Our ultimate goal is to establish a design principle of a controller in order to achieve natural human-like locomotion. We suggest dynamical movement primitives as a CPG of a biped robot, an approach we have previously proposed for learning and encoding complex human movements. Demonstrated trajectories are learned through the movement primitives by locally weighted regression, and the frequency of the learned trajectories is adjusted automatically by a novel frequency adaptation algorithm based on phase resetting and entrainment of oscillators. Numerical simulations demonstrate the effectiveness of the proposed locomotion controller.

am

link (url) [BibTex]

link (url) [BibTex]


no image
Movement planning and imitation by shaping nonlinear attractors

Schaal, S.

In Proceedings of the 12th Yale Workshop on Adaptive and Learning Systems, Yale University, New Haven, CT, 2003, clmc (inproceedings)

Abstract
Given the continuous stream of movements that biological systems exhibit in their daily activities, an account for such versatility and creativity has to assume that movement sequences consist of segments, executed either in sequence or with partial or complete overlap. Therefore, a fundamental question that has pervaded research in motor control both in artificial and biological systems revolves around identifying movement primitives (a.k.a. units of actions, basis behaviors, motor schemas, etc.). What are the fundamental building blocks that are strung together, adapted to, and created for ever new behaviors? This paper summarizes results that led to the hypothesis of Dynamic Movement Primitives (DMP). DMPs are units of action that are formalized as stable nonlinear attractor systems. They are useful for autonomous robotics as they are highly flexible in creating complex rhythmic (e.g., locomotion) and discrete (e.g., a tennis swing) behaviors that can quickly be adapted to the inevitable perturbations of a dy-namically changing, stochastic environment. Moreover, DMPs provide a formal framework that also lends itself to investigations in computational neuroscience. A recent finding that allows creating DMPs with the help of well-understood statistical learning methods has elevated DMPs from a more heuristic to a principled modeling approach, and, moreover, created a new foundation for imitation learning. Theoretical insights, evaluations on a humanoid robot, and behavioral and brain imaging data will serve to outline the framework of DMPs for a general approach to motor control and imitation in robotics and biology.

am

link (url) [BibTex]

link (url) [BibTex]