Header logo is


2014


Probabilistic Progress Bars
Probabilistic Progress Bars

Kiefel, M., Schuler, C., Hennig, P.

In Conference on Pattern Recognition (GCPR), 8753, pages: 331-341, Lecture Notes in Computer Science, (Editors: Jiang, X., Hornegger, J., and Koch, R.), Springer, GCPR, September 2014 (inproceedings)

Abstract
Predicting the time at which the integral over a stochastic process reaches a target level is a value of interest in many applications. Often, such computations have to be made at low cost, in real time. As an intuitive example that captures many features of this problem class, we choose progress bars, a ubiquitous element of computer user interfaces. These predictors are usually based on simple point estimators, with no error modelling. This leads to fluctuating behaviour confusing to the user. It also does not provide a distribution prediction (risk values), which are crucial for many other application areas. We construct and empirically evaluate a fast, constant cost algorithm using a Gauss-Markov process model which provides more information to the user.

ei ps pn

website+code pdf DOI [BibTex]

2014


website+code pdf DOI [BibTex]


Probabilistic Solutions to Differential Equations and their Application to Riemannian Statistics
Probabilistic Solutions to Differential Equations and their Application to Riemannian Statistics

Hennig, P., Hauberg, S.

In Proceedings of the 17th International Conference on Artificial Intelligence and Statistics, 33, pages: 347-355, JMLR: Workshop and Conference Proceedings, (Editors: S Kaski and J Corander), Microtome Publishing, Brookline, MA, AISTATS, April 2014 (inproceedings)

Abstract
We study a probabilistic numerical method for the solution of both boundary and initial value problems that returns a joint Gaussian process posterior over the solution. Such methods have concrete value in the statistics on Riemannian manifolds, where non-analytic ordinary differential equations are involved in virtually all computations. The probabilistic formulation permits marginalising the uncertainty of the numerical solution such that statistics are less sensitive to inaccuracies. This leads to new Riemannian algorithms for mean value computations and principal geodesic analysis. Marginalisation also means results can be less precise than point estimates, enabling a noticeable speed-up over the state of the art. Our approach is an argument for a wider point that uncertainty caused by numerical calculations should be tracked throughout the pipeline of machine learning algorithms.

ei ps pn

pdf Youtube Supplements Project page link (url) [BibTex]

pdf Youtube Supplements Project page link (url) [BibTex]


no image
Probabilistic ODE Solvers with Runge-Kutta Means

Schober, M., Duvenaud, D., Hennig, P.

In Advances in Neural Information Processing Systems 27, pages: 739-747, (Editors: Z. Ghahramani, M. Welling, C. Cortes, N.D. Lawrence and K.Q. Weinberger), Curran Associates, Inc., 28th Annual Conference on Neural Information Processing Systems (NIPS), 2014 (inproceedings)

ei pn

Web link (url) [BibTex]

Web link (url) [BibTex]


no image
Active Learning of Linear Embeddings for Gaussian Processes

Garnett, R., Osborne, M., Hennig, P.

In Proceedings of the 30th Conference on Uncertainty in Artificial Intelligence, pages: 230-239, (Editors: NL Zhang and J Tian), AUAI Press , Corvallis, Oregon, UAI2014, 2014, another link: http://arxiv.org/abs/1310.6740 (inproceedings)

ei pn

PDF Web [BibTex]

PDF Web [BibTex]


no image
Probabilistic Shortest Path Tractography in DTI Using Gaussian Process ODE Solvers

Schober, M., Kasenburg, N., Feragen, A., Hennig, P., Hauberg, S.

In Medical Image Computing and Computer-Assisted Intervention – MICCAI 2014, Lecture Notes in Computer Science Vol. 8675, pages: 265-272, (Editors: P. Golland, N. Hata, C. Barillot, J. Hornegger and R. Howe), Springer, Heidelberg, MICCAI, 2014 (inproceedings)

ei pn

DOI [BibTex]

DOI [BibTex]


no image
Sampling for Inference in Probabilistic Models with Fast Bayesian Quadrature

Gunter, T., Osborne, M., Garnett, R., Hennig, P., Roberts, S.

In Advances in Neural Information Processing Systems 27, pages: 2789-2797, (Editors: Z. Ghahramani, M. Welling, C. Cortes, N.D. Lawrence and K.Q. Weinberger), Curran Associates, Inc., 28th Annual Conference on Neural Information Processing Systems (NIPS), 2014 (inproceedings)

ei pn

Web link (url) [BibTex]

Web link (url) [BibTex]


{Detecting magnetic flux distributions in superconductors with polarized x-rays}
Detecting magnetic flux distributions in superconductors with polarized x-rays

Stahl, C., Audehm, P., Gräfe, J., Ruoß, S., Weigand, M., Schmidt, M., Treiber, S., Bechtel, M., Goering, E., Schütz, G., Albrecht, J.

{Physical Review B}, 90(10), American Physical Society, Woodbury, NY, 2014 (article)

Abstract
The magnetic flux distribution arising from a high-Tc superconductor is detected and visualized using polarized x rays. Therefore, we introduce a sensor layer, namely, an amorphous, soft-magnetic Co40Fe40B20 cover layer, providing a large x-ray magnetic circular dichroism (XMCD). Temperature-dependent XMCD spectroscopy on the magnetic layer has been performed. Exploiting the temperature dependence of the critical current density of the superconductor we find a quantitative correlation between the XMCD signal and the in-plane stray field of the superconductor. Magneto-optical Kerr effect experiments on the sensor layer can simulate the stray field of the superconductor and hence verify the correlation. We show that the XMCD contrast in the sensor layer corresponds to the in-plane magnetic flux distribution of the superconductor and can hence be used to image magnetic structures in superconductors.

mms

DOI [BibTex]

DOI [BibTex]


no image
Incremental Local Gaussian Regression

Meier, F., Hennig, P., Schaal, S.

In Advances in Neural Information Processing Systems 27, pages: 972-980, (Editors: Z. Ghahramani, M. Welling, C. Cortes, N.D. Lawrence and K.Q. Weinberger), 28th Annual Conference on Neural Information Processing Systems (NIPS), 2014, clmc (inproceedings)

am ei pn

PDF link (url) [BibTex]

PDF link (url) [BibTex]


{Application of magneto-optical Kerr effect to first-order reversal curve measurements}
Application of magneto-optical Kerr effect to first-order reversal curve measurements

Gräfe, J., Schmidt, M., Audehm, P., Schütz, G., Goering, E.

{Review of Scientific Instruments}, 85, American Institute of Physics, Woodbury, N.Y. [etc.], 2014 (article)

Abstract
First-order reversal curves (FORC) are a powerful method for magnetic sample characterization, separating all magnetic states of an investigated system according to their coercivity and internal magnetic interactions. A major drawback of using measurement techniques like VSM or SQUID, typically applied for FORC acquisition, is the long measurement time, limiting the resolution and the number of measurements due to time constraints. Faster techniques like MOKE result in problems regarding measurement stability over the curse of the acquisition of many minor loops, due to drift and non-absolute magnetization values. Here, we present an approach using a specialized field shape providing two anchor points for each minor loop for applying the magneto-optical Kerr effect (MOKE) technique to FORC measurements. This results in a high field resolution while keeping the total acquisition time to only a few minutes. MOKE FORC measurements are exemplarily applied to a simple permalloy film, an exchange-bias system, and a Gd/Fe multilayer system with perpendicular magnetic anisotropy, showcasing the versatility of the method.

mms

DOI Project Page [BibTex]

DOI Project Page [BibTex]


no image
Efficient Bayesian Local Model Learning for Control

Meier, F., Hennig, P., Schaal, S.

In Proceedings of the IEEE International Conference on Intelligent Robots and Systems, pages: 2244 - 2249, IROS, 2014, clmc (inproceedings)

Abstract
Model-based control is essential for compliant controland force control in many modern complex robots, like humanoidor disaster robots. Due to many unknown and hard tomodel nonlinearities, analytical models of such robots are oftenonly very rough approximations. However, modern optimizationcontrollers frequently depend on reasonably accurate models,and degrade greatly in robustness and performance if modelerrors are too large. For a long time, machine learning hasbeen expected to provide automatic empirical model synthesis,yet so far, research has only generated feasibility studies butno learning algorithms that run reliably on complex robots.In this paper, we combine two promising worlds of regressiontechniques to generate a more powerful regression learningsystem. On the one hand, locally weighted regression techniquesare computationally efficient, but hard to tune due to avariety of data dependent meta-parameters. On the other hand,Bayesian regression has rather automatic and robust methods toset learning parameters, but becomes quickly computationallyinfeasible for big and high-dimensional data sets. By reducingthe complexity of Bayesian regression in the spirit of local modellearning through variational approximations, we arrive at anovel algorithm that is computationally efficient and easy toinitialize for robust learning. Evaluations on several datasetsdemonstrate very good learning performance and the potentialfor a general regression learning tool for robotics.

am ei pn

PDF link (url) DOI [BibTex]

PDF link (url) DOI [BibTex]


no image
Efflorescence upon humidification? X-ray microspectroscopic in situ observation of changes in aerosol microstructure and phase state upon hydration

Pöhlker, C., Saturno, J., Krüger, M. L., Förster, J. D., Weigand, M., Wiedemann, K. T., Bechtel, M., Artaxo, P., Andreae, M. O.

{Geophysical Research Letters}, 41(10):3681-3689, American Geophysical Union, Washington, D.C., 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Investigation of cellular microstructure and enhanced coercivity in sputtered Sm2(CoCuFeZr)17 film

Bhatt, R., Bhatt, P., Schütz, G.

{Journal of Applied Physics}, 115(10), American Institute of Physics, New York, NY, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Spatial degradation mapping and component-wise degradation tracking in polymer-fullerene blends

Pedersen, E. B. B., Tromholt, T., Madsen, M. V., Böttiger, A. P. L., Weigand, M., Krebs, F. C., Andreasen, J. W.

{Journal of Materials Chemistry C}, 2(26):5176-5182, Royal Society of Chemistry, Cambridge, UK, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Vortex core reversal due to spin wave interference

Bauer, H. G., Sproll, M., Back, C. H., Woltersdorf, G.

{Physical Review Letters}, 112, American Physical Society, Woodbury, N.Y., 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Increasing the sensor performance using Au modified high temperature superconducting YBa2Cu3O7-delta thin films

Katzer, C., Stahl, C., Michalowski, P., Treiber, S., Westernhausen, M., Schmidl, F., Seidel, P., Schütz, G., Albrecht, J.

In 507, IOP Pub., Genova, Italy, 2014 (inproceedings)

mms

DOI [BibTex]

DOI [BibTex]


no image
Understanding the adsorption mechanism of noble gases Kr and Xe in CPO-27-Ni, CPO-27-Mg, and ZIF-8

Magdysyuk, O. V., Adams, F., Liermann, H., Spanopoulos, I., Trikalitis, P. N., Hirscher, M., Morris, R. E., Duncan, M. J., McCormick, L. J., Dinnebier, R. E.

{Physical Chemistry Chemical Physics}, 16(43):23908-23914, Royal Society of Chemistry, Cambridge, England, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Unidirectional sub-100-ps magnetic vortex core reversal

Noske, M., Gangwar, A., Stoll, H., Kammerer, M., Sproll, M., Dieterle, G., Weigand, M., Fähnle, M., Woltersdorf, G., Back, C. H., Schütz, G.

{Physical Review B}, 90(10), American Physical Society, Woodbury, NY, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Domain wall transformations and hopping in La0.7Sr0.3MnO3 nanostructures imaged with high resolution x-ray magnetic microscopy

Finizio, S., Foerster, M., Krüger, B., Vaz, C. A. F., Miyawaki, T., Mawass, M. A., Pena, L., Méchin, L., Hühn, S., Moshnyaga, V., Büttner, F., Bisig, A., Le Guyader, L., El Moussaoui, S., Valencia, S., Kronast, F., Eisebitt, S., Kläui, M.

{Journal of Physics: Condensed Matter}, 26(45), IOP Publishing, Bristol, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Tunable eigenmodes of coupled magnetic vortex oscillators

Hänze, M., Adolff, C. F., Weigand, M., Meier, G.

{Applied Physics Letters}, 104(18), American Institute of Physics, Melville, NY, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Domain wall pinning in ultra-narrow electromigrated break junctions

Reeve, R. M., Loescher, A., Mawass, M.-A., Hoffmann-Vogel, R., Kläui, M.

{Journal of Physics: Condensed Matter}, 26(47), IOP Publishing, Bristol, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Highly effective hydrogen isotope separation in nanoporous metal-organic framworks with open metal sites: Direct measurement and theoretical analysis

Oh, H., Savchenko, I., Mavrandonakis, A., Heine, T., Hirscher, M.

{ACS Nano}, 8(1):761-770, American Chemical Society, Washington, DC, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Stabilization of the dissipation-free current transport in inhomogeneous MgB2 thin films

Treiber, S., Stahl, C., Schütz, G., Soltan, S., Albrecht, J.

{Physica C}, 506, pages: 1-5, North-Holland, Amsterdam, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Thermal conductivity of mechanically joined semiconducting/metal nanomembrane superlattices

Grimm, D., Wilson, R. B., Teshome, B., Gorantla, S., Rümmeli, M. H., Bublat, T., Zallo, E., Li, G., Cahill, D. G., Schmidt, O. G.

{Nano Letters}, 14(5):2387-2393, American Chemical Society, Washington, DC, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Role of the sample boundaries in the problem of dissipative magnetization dynamics

Fähnle, M., Slavin, A., Hertel, R.

{Journal of Magnetism and Magnetic Materials}, 360, pages: 126-130, Elsevier, Amsterdam, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Using magnetic coupling in bilayers of superconducting YBCO and soft-magnetic CoFeB to map supercurrent flow

Stahl, C., Walker, P., Treiber, S., Christiani, G., Schütz, G., Albrecht, J.

{EPL}, 106(2), 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Element-specific depth profile of magnetism and stoichiometry at the La0.67Sr0.33MnO3/BiFeO3 interface

Bertinshaw, J., Brück, S., Lott, D., Fritzsche, H., Khaydukov, Y., Soltwedel, O., Keller, T., Goering, E., Audehm, P., Cortie, D. L., Hutchison, W. D., Ramasse, Q. M., Arredondo, M., Maran, R., Nagarajan, V., Klose, F., Ulrich, C.

{Physical Review B}, 90(4), American Physical Society, Woodbury, NY, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Magnetic reflectometry of heterostructures (Topical Review)

Macke, S., Goering, E.

{Journal of Physics: Condensed Matter}, 26(36), IOP Publishing, Bristol, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Optimizing the fabrication of diffractive optical elements using a focused ion beam system

Vijayakumar, A., Eigenthaler, U., Keskinbora, K., Sridharan, G. M., Pramitha, V., Hirscher, M., Spatz, J. P., Bhattacharya, S.

{Proceedings of SPIE}, 9130, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Direct observation of internal vortex domain-wall dynamics

Stein, F.-U., Bocklage, L., Weigand, M., Meier, G.

{Physical Review B}, 89(2), American Physical Society, Woodbury, NY, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Synchronous precessional motion of multiple domain walls in a ferromagnetic nanowire by perpendicular field pulses

Kim, J., Mawass, M., Bisig, A., Krüger, B., Reeve, R. M., Schulz, T., Büttner, F., Yoon, J., You, C., Weigand, M., Stoll, H., Schütz, G., Swagten, H. J. M., Koopmans, B., Eisebitt, S., Kläui, M.

{Nature Communications}, 5, Nature Publishing Group, London, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Magnetic field distribution and characteristic fields of the vortex lattice for a clean superconducting niobium sample in an external field applied along a three-fold axis

Yaouanc, A., Maisuradze, A., Nakai, N., Machida, K., Khasanov, R., Amato, A., Biswas, P. K., Baines, C., Herlach, D., Henes, Rolf, Keppler, P., Keller, H.

{Physical Review B}, 89(18), American Physical Society, Woodbury, NY, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Experimental assessment of Physical upper limit for hydrogen storage capacity at 20 K in densified MIL-101 monoliths

Oh, H., Lupu, D., Blanita, G., Hirscher, M.

{RSC Advances}, 4(6):2648-2651, Royal Society of Chemistry, Cambridge, UK, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Strengthening zones in the Co matrix of WC-Co cemented carbides

Konyashin, I., Lachmann, F., Ries, B., Mazilkin, A. A., Straumal, B. B., Kübel, C., Llanes, L., Baretzky, B.

{Scripta Materialia}, 83, pages: 17-20, Pergamon, Tarrytown, NY, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Multilayer Fresnel zone plates for high energy radiation resolve 21 nm features at 1.2 keV

Keskinbora, K., Robisch, A., Mayer, M., Sanli, U., Grévent, C., Wolter, C., Weigand, M., Szeghalmi, A., Knez, M., Salditt, T., Schütz, G.

{Optics Express}, 22(15):18440-18453, Optical Society of America, Washington, DC, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Interplay of linker functionalization and hydrogen adsorption in the metal-organic framework MIL-101

Szilágyi, P. A., Weinrauch, I., Oh, H., Hirscher, M., Juan-Alcaniz, J., Serra-Crespo, P., de Respinis, M., Trzesniewski, B. J., Kapteijn, F., Geerlings, H., Gascon, J., Dam, B., Grzech, A., van de Krol, R.

{The Journal of Physical Chemistry C}, 118(34):19572-19579, American Chemical Society, Washington DC, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Efficient focusing of 8 keV X-rays with multilayer Fresnel zone plates fabricated by atomic layer deposition and focused ion beam milling. Erratum

Mayer, M., Keskinbora, K., Grévent, C., Szeghalmi, A., Knez, M., Weigand, M., Snigirev, A., Snigireva, I., Schütz, G.

{Journal of Synchrotron Radiation}, 640, pages: 640-640, Published for the International Union of Crystallography by Munksgaard, Copenhagen, Denmark, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Low-amplitude magnetic vortex core reversal by non-linear interaction between azimuthal spin waves and the vortex gyromode

Sproll, M., Noske, M., Bauer, H., Kammerer, M., Gangwar, A., Dieterle, G., Weigand, M., Stoll, H., Woltersdorf, G., Back, C. H., Schütz, G.

{Applied Physics Letters}, 104(1), American Institute of Physics, Melville, NY, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
The local magnetic properties of [MnIII6 CrIII]3+ and [FeIII6 CrIII]3+ single-molecule magnets deposited on surfaces studied by spin-polarized photoemission and XMCD with circularly polarized synchrotron radiation

Heinzmann, U., Helmstedt, A., Dohmeier, N., Müller, N., Gryzia, A., Brechling, A., Hoeke, V., Krickemeyer, E., Glaser, T., Fonin, M., Bouvron, S., Leicht, P., Tietze, T., Goering, E., Kuepper, K.

{Journal of Physics: Conference Series}, 488(13), IOP Publishing, Bristol, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
A fluorene based covalent triazine framework with high CO2 and H2 capture and storage capacities

Hug, S., Mesch, M. B., Oh, H., Popp, N., Hirscher, M., Senker, J., Lotsch, B. V.

{Journal of Materials Chemistry A}, 2(16):5928-5936, Royal Society of Chemistry, Cambridge, UK, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Ab-initio calculations and atomistic calculations on the magnetoelectric effects in metallic nanostructures

Fähnle, M., Subkow, S.

{Physica Status Solidi C}, 11(2):185-191, Wiley-VCH, Weinheim, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Role of electron-magnon scatterings in ultrafast demagnetization

Haag, M., Illg, C., Fähnle, M.

{Physical Review B}, 90(1), American Physical Society, Woodbury, NY, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Element specific monolayer depth profiling

Macke, S., Radi, A., Hamann-Borrero, J. E., Verna, A., Bluschke, M., Brück, S., Goering, E., Sutarto, R., He, F., Cristiani, G., Wu, M., Benckiser, E., Habermeier, H., Logvenov, G., Gauquelin, N., Botton, G. A., Kajdos, A. P., Stemmer, S., Sawatzky, G. A., Haverkort, M. W., Keimer, B., Hinkov, V.

{Advanced Materials}, 26(38):6554-6559, Wiley VCH, Weinheim, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Local modification of the magnetic vortex-core velocity by gallium implantation

Langner, H. H., Vogel, A., Beyersdorff, B., Weigand, M., Frömter, R., Oepen, H. P., Meier, G.

{Journal of Applied Physcis}, (10), American Institute of Physics, New York, NY, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Influence of magnetic fields on spin-mixing in transition metals

Haag, M., Illg, C., Fähnle, M.

{Physical Review B}, 90(13), American Physical Society, Woodbury, NY, 2014 (article)

mms

DOI [BibTex]

DOI [BibTex]

2010


no image
Using an Infinite Von Mises-Fisher Mixture Model to Cluster Treatment Beam Directions in External Radiation Therapy

Bangert, M., Hennig, P., Oelfke, U.

In pages: 746-751 , (Editors: Draghici, S. , T.M. Khoshgoftaar, V. Palade, W. Pedrycz, M.A. Wani, X. Zhu), IEEE, Piscataway, NJ, USA, Ninth International Conference on Machine Learning and Applications (ICMLA), December 2010 (inproceedings)

Abstract
We present a method for fully automated selection of treatment beam ensembles for external radiation therapy. We reformulate the beam angle selection problem as a clustering problem of locally ideal beam orientations distributed on the unit sphere. For this purpose we construct an infinite mixture of von Mises-Fisher distributions, which is suited in general for density estimation from data on the D-dimensional sphere. Using a nonparametric Dirichlet process prior, our model infers probability distributions over both the number of clusters and their parameter values. We describe an efficient Markov chain Monte Carlo inference algorithm for posterior inference from experimental data in this model. The performance of the suggested beam angle selection framework is illustrated for one intra-cranial, pancreas, and prostate case each. The infinite von Mises-Fisher mixture model (iMFMM) creates between 18 and 32 clusters, depending on the patient anatomy. This suggests to use the iMFMM directly for beam ensemble selection in robotic radio surgery, or to generate low-dimensional input for both subsequent optimization of trajectories for arc therapy and beam ensemble selection for conventional radiation therapy.

ei pn

Web DOI [BibTex]

2010


Web DOI [BibTex]


no image
Coherent Inference on Optimal Play in Game Trees

Hennig, P., Stern, D., Graepel, T.

In JMLR Workshop and Conference Proceedings Volume 9: AISTATS 2010, pages: 326-333, (Editors: Teh, Y.W. , M. Titterington ), JMLR, Cambridge, MA, USA, Thirteenth International Conference on Artificial Intelligence and Statistics, May 2010 (inproceedings)

Abstract
Round-based games are an instance of discrete planning problems. Some of the best contemporary game tree search algorithms use random roll-outs as data. Relying on a good policy, they learn on-policy values by propagating information upwards in the tree, but not between sibling nodes. Here, we present a generative model and a corresponding approximate message passing scheme for inference on the optimal, off-policy value of nodes in smooth AND/OR trees, given random roll-outs. The crucial insight is that the distribution of values in game trees is not completely arbitrary. We define a generative model of the on-policy values using a latent score for each state, representing the value under the random roll-out policy. Inference on the values under the optimal policy separates into an inductive, pre-data step and a deductive, post-data part. Both can be solved approximately with Expectation Propagation, allowing off-policy value inference for any node in the (exponentially big) tree in linear time.

ei pn

PDF Web [BibTex]

PDF Web [BibTex]


no image
Dzyaloshinskii-Moriya interactions in systems with fabrication induced strain gradients: ab-initio study

Beck, P., Fähnle, M

{Journal of Magnetism and Magnetic Materials}, 322, pages: 3701-3703, 2010 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
On the nature of displacement bursts during nanoindentation of ultrathin Ni films on sapphire

Rabkin, E., Deuschle, J. K., Baretzky, B.

{Acta Materialia}, 58, pages: 1589-1598, 2010 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Nanospheres generate out-of-plane magnetization

Amaladass, E., Ludescher, B., Schütz, G., Tyliszczak, T., Lee, M., Eimüller, T.

{Journal of Applied Physics}, 107, 2010 (article)

mms

DOI [BibTex]

DOI [BibTex]


no image
Absence of element specific ferromagnetism in Co doped ZnO investigated by soft X-ray resonant reflectivity

Goering, E., Brück, S., Tietze, T., Jakob, G., Gacic, M., Adrian, H.

In 200, Glasgow, Scotland, 2010 (inproceedings)

mms

DOI [BibTex]

DOI [BibTex]