5,313 research outputs found
Coherent X-ray Diffraction Imaging
No full textFor centuries, lens-based microscopy, such as optical, phase-contrast, fluorescence, confocal, and electron microscopy, has played an important role in the evolution of modern science and technology. In 1999, a novel form of microscopy, i.e., coherent diffraction imaging (also termed coherent diffraction microscopy or lensless imaging), was developed and transformed our conventional view of microscopy, in which the diffraction pattern of a noncrystalline specimen or a nanocrystal was first measured and then directly phased to obtain a high-resolution image. The well-known phase problem was solved by combining the oversampling method with iterative algorithms. In this paper, we will briefly discuss the principle of coherent diffraction imaging, present various implementation schemes of this imaging modality, and illustrate its broad applications in materials science, nanoscience, and biology. As coherent X-ray sources such as high harmonic generation and X-ray free-electron lasers are presently under rapid development worldwide, coherent diffraction imaging can potentially be applied to perform high-resolution imaging of materials/nanoscience and biological specimens at the femtosecond time scale.X115360sciescopu
Teacher-apprentices RL (TARL): leveraging complex policy distribution through generative adversarial hypernetwork in reinforcement learning
No full textTypically, a Reinforcement Learning (RL) algorithm focuses in learning a single deployable policy as the end product. Depending on the initialization methods and seed randomization, learning a single policy could possibly leads to convergence to different local optima across different runs, especially when the algorithm is sensitive to hyper-parameter tuning. Motivated by the capability of Generative Adversarial Networks (GANs) in learning complex data manifold, the adversarial training procedure could be utilized to learn a population of good-performing policies instead. We extend the teacher-student methodology observed in the Knowledge Distillation field in typical deep neural network prediction tasks to RL paradigm. Instead of learning a single compressed student network, an adversarially-trained generative model (hypernetwork) is learned to output network weights of a population of good-performing policy networks, representing a school of apprentices. Our proposed framework, named Teacher-Apprentices RL (TARL), is modular and could be used in conjunction with many existing RL algorithms. We illustrate the performance gain and improved robustness by combining TARL with various types of RL algorithms, including direct policy search Cross-Entropy Method, Q-learning, Actor-Critic, and policy gradient-based methods.Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.Interactive Intelligenc
XMM-Newton spectroscopy of an X-ray selected sample of RL AGNs
No full textThis paper presents the X-ray spectroscopy of an X-ray selected sample
of 25 radio-loud (RL) AGNs extracted from the XMM-Newton Bright Serendipitous
Survey (XBSS). The main goal of the work is to assess and study the origin
of the X-ray spectral differences usually observed between radio-loud and radio-quiet (RQ) AGNs. To this end, a comparison sample of 53 RQ AGNs
has been also extracted from the same XBSS sample and
studied together with the sample of RL AGNs.
Since there are many claims in the literature that RL AGNs have, on average, a flatter spectral
index when compared to the RQ AGNs, we have focused the analysis on the distribution of
the X-ray spectral indices of the power-law component that models the large majority of the
spectra in both samples.
We find that the mean X-ray energy spectral index is very similar
in the 2 samples and close to . However, the intrinsic distribution of the
spectral indices is significantly broader in the sample of RL AGNs. In order to investigate
the origin of this difference, we have divided the RL AGNs into blazars (i.e. BL Lac objects and
FSRQs) and “non-blazars” (i.e. radiogalaxies and SSRQs), on the basis of the available
optical and radio information. Although the number of sources is small, we find strong
evidence that the broad distribution observed in the RL AGN sample is mainly due to the
presence of the blazars. Furthermore, within the blazar class we have found a link between
the X-ray spectral index and the value of the radio-to-X-ray spectral index ()
suggesting that the observed X-ray emission is directly connected to the
emission of the relativistic jet. This trend is not observed among
the “non-blazars” RL AGNs. This favours the hypothesis that, in these latter sources, the
X-ray emission is not significantly influenced by the jet emission and it has probably an
origin similar to the RQ AGNs. Overall, the results presented here indicate
that the observed distribution of the X-ray spectral indices in a given sample of RL AGNs is
strongly dependent on the amount of relativistic beaming present in the selected sources, i.e.
on the relative fraction of blazars and “non-blazars”
BADDr: Bayes-Adaptive Deep Dropout RL for POMDPs
No full textWhile reinforcement learning (RL) has made great advances in scalability, exploration and partial observability are still active research topics. In contrast, Bayesian RL (BRL) provides a principled answer to both state estimation and the exploration-exploitation trade-off, but struggles to scale. To tackle this challenge, BRL frameworks with various prior assumptions have been proposed, with varied success. This work presents a representation-agnostic formulation of BRL under partially observability, unifying the previous models under one theoretical umbrella. To demonstrate its practical significance we also propose a novel derivation, Bayes-Adaptive Deep Dropout rl (BADDr), based on dropout networks. Under this parameterization, in contrast to previous work, the belief over the state and dynamics is a more scalable inference problem. We choose actions through Monte-Carlo tree search and empirically show that our method is competitive with state-of-the-art BRL methods on small domains while being able to solve much larger ones.Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.Interactive Intelligenc
Refined Risk Management in Safe Reinforcement Learning with a Distributional Safety Critic
No full textSafety is critical to broadening the real-world use of reinforcement learning (RL). Modeling the safety aspects using a safety-cost signal separate from the reward is becoming standard practice, since it avoids the problem of finding a good balance between safety and performance. However, the total safety-cost distribution of different trajectories is still largely unexplored. In this paper, we propose an actor critic method for safe RL that uses an implicit quantile network to approximate the distribution of accumulated safety-costs. Using an accurate estimate of the distribution of accumulated safetycosts, in particular of the upper tail of the distribution, greatly improves the performance of riskaverse RL agents. The empirical analysis shows that our method achieves good risk control in complex safety-constrained environments.AlgorithmicsIntelligent Electrical Power Grid
Changes in wheat germination following gamma-ray irradiation: an in vivo electronic paramagnetic resonance spin-probe study
No full textEmbryos excised from wheat (Triticum aestivum) grains following gamma-ray irradiation at different doses were analyzed on membrane permeability by electron paramagnetic resonance (EPR) technique with 4-oxo-2, 2, 6, 6-tetramethyl-1-piperidinyloxy (TEMPONE) as spin probe to acquire an EPR spectrum. The broadening agent ferricyanide added leads to changes in the high-field region of the EPR spectrum, which reflects differences in membrane permeability. R-value, defined as the ratio of water (W) to lipid (L) component in height in the high-field region of the EPR spectrum, symbolizes membrane permeability for a given sample. The R-values corresponding to a certain dose treatment of grains displayed a definitive distribution pattern. A unit row vector with 20 components was used to describe the R-value distribution pattern for a given treatment. The transaction angle between vectors corresponding to grains irradiated and unirradiated, theta, was used as quantitative index for membrane permeability changes following gamma-ray irradiation. gamma-Ray irradiated grains germinated at low rates, and the regression equation of germination rate as a function of the irradiation dose is: Germination Rate (%) = 94.8 exp[- 0.264 x Irradiation Dose (kGy)] (r(2) = 0.991, P < 0.001). Embryos excised from grains following gamma-ray irradiation show increases in theta values with irradiation dose. The theta value is negatively linearly correlated with the germination rate. It suggests that gamma-ray irradiation leading to increases in membrane permeability is consistent with that leading to low germination rate of grains. The introduction to vector analysis method on membrane permeability changes in this study is very practical. (C) 2000 Elsevier Science B.V. All rights reserved.Plant SciencesEnvironmental SciencesSCI(E)6ARTICLE3219-2254
qgym: A Gym for Training and Benchmarking RL-Based Quantum Compilation
No full textCompiling a quantum circuit for specific quantum hardware is a challenging task. Moreover, current quantum computers have severe hardware limitations. To make the most use of the limited resources, the compilation process should be optimized. To improve currents methods, Reinforcement Learning (RL), a technique in which an agent interacts with an environment to learn complex policies to attain a specific goal, can be used. In this work, we present qgym, a software framework derived from the OpenAI gym, together with environments that are specifically tailored towards quantum compilation. The goal of qgym is to connect the research fields of Artificial Intelligence (AI) with quantum compilation by abstracting parts of the process that are irrelevant to either domain. It can be used to train and benchmark RL agents and algorithms in highly customizable environments.Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.Quantum Circuit Architectures and Technolog
SYNTHESIS OF TRIMETHYLSILYLATED GERMANOCENES - X-RAY STRUCTURE OF AND STERIC EFFECTS IN HEXAKIS(TRIMETHYLSILYL)GERMANOCENE
No full textJutzi P, SCHLUTER E, HURSTHOUSE MB, ARIF AM, SHORT RL. SYNTHESIS OF TRIMETHYLSILYLATED GERMANOCENES - X-RAY STRUCTURE OF AND STERIC EFFECTS IN HEXAKIS(TRIMETHYLSILYL)GERMANOCENE. JOURNAL OF ORGANOMETALLIC CHEMISTRY. 1986;299(3):285-295
Influence-Augmented Local Simulators: a Scalable Solution for Fast Deep RL in Large Networked Systems
No full textLearning effective policies for real-world problems is still an open challenge for the field of reinforcement learning (RL). The main limitation being the amount of data needed and the pace at which that data can be obtained. In this paper, we study how to build lightweight simulators of complicated systems that can run sufficiently fast for deep RL to be applicable. We focus on domains where agents interact with a reduced portion of a larger environment while still being affected by the global dynamics. Our method combines the use of local simulators with learned models that mimic the influence of the global system. The experiments reveal that incorporating this idea into the deep RL workflow can considerably accelerate the training process and presents several opportunities for the future.Interactive IntelligenceAlgorithmic
LONG-TERM NATURAL HISTORY OF FOCAL CONFLUENT FIBROSIS (FCF) IN THE CIRRHOTIC LIVER: A CT SERIAL ANALYSIS
No full text- …
