Unconstrained monocular 3D human pose estimation by action detection and cross-modality regression forest

This work addresses the challenging problem of unconstrained 3D human pose estimation (HPE) from a novel perspective. Existing approaches struggle to operate in realistic applications, mainly due to their scene-dependent priors, such as background segmentation and multi-camera network, which restrict their use in unconstrained environments. We therfore present a framework which applies action detection and 2D pose estimation techniques to infer 3D poses in an unconstrained video. Action detection offers spatiotemporal priors to 3D human pose estimation by both recognising and localising actions in space-time. Instead of holistic features, e.g. silhouettes, we leverage the flexibility of deformable part model to detect 2D body parts as a feature to estimate 3D poses. A new unconstrained pose dataset has been collected to justify the feasibility of our method, which demonstrated promising results, significantly outperforming the relevant state-of-the-arts. © 2013 IEEE.

Free energy surface reconstruction from umbrella samples using Gaussian process regression

We demonstrate how the Gaussian process regression approach can be used to efficiently reconstruct free energy surfaces from umbrella sampling simulations. By making a prior assumption of smoothness and taking account of the sampling noise in a consistent fashion, we achieve a significant improvement in accuracy over the state of the art in two or more dimensions or, equivalently, a significant cost reduction to obtain the free energy surface within a prescribed tolerance in both regimes of spatially sparse data and short sampling trajectories. Stemming from its Bayesian interpretation the method provides meaningful error bars without significant additional computation. A software implementation is made available on www.libatoms.org.

Free energy surface reconstruction from umbrella samples using Gaussian process regression II: Multiple collective variables

We demonstrate how a prior assumption of smoothness can be used to enhance the reconstruction of free energy profiles from multiple umbrella sampling simulations using the Bayesian Gaussian process regression approach. The method we derive allows the concurrent use of histograms and free energy gradients and can easily be extended to include further data. In Part I we review the necessary theory and test the method for one collective variable. We demonstrate improved performance with respect to the weighted histogram analysis method and obtain meaningful error bars without any significant additional computation. In Part II we consider the case of multiple collective variables and compare to a reconstruction using least squares fitting of radial basis functions. We find substantial improvements in the regimes of spatially sparse data or short sampling trajectories. A software implementation is made available on www.libatoms.org.

Automatic construction and natural-language description of nonparametric regression models

Copyright © 2014, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved. This paper presents the beginnings of an automatic statistician, focusing on regression problems. Our system explores an open-ended space of statistical models to discover a good explanation of a data set, and then produces a detailed report with figures and natural- language text. Our approach treats unknown regression functions non- parametrically using Gaussian processes, which has two important consequences. First, Gaussian processes can model functions in terms of high-level properties (e.g. smoothness, trends, periodicity, changepoints). Taken together with the compositional structure of our language of models this allows us to automatically describe functions in simple terms. Second, the use of flexible nonparametric models and a rich language for composing them in an open-ended manner also results in state- of-the-art extrapolation performance evaluated over 13 real time series data sets from various domains.

First principles interatomic potential for tungsten based on Gaussian process regression

An accurate description of atomic interactions, such as that provided by first principles quantum mechanics, is fundamental to realistic prediction of the properties that govern plasticity, fracture or crack propagation in metals. However, the computational complexity associated with modern schemes explicitly based on quantum mechanics limits their applications to systems of a few hundreds of atoms at most. This thesis investigates the application of the Gaussian Approximation Potential (GAP) scheme to atomistic modelling of tungsten - a bcc transition metal which exhibits a brittle-to-ductile transition and whose plasticity behaviour is controlled by the properties of $\frac{1}{2} \langle 111 \rangle$ screw dislocations. We apply Gaussian process regression to interpolate the quantum-mechanical (QM) potential energy surface from a set of points in atomic configuration space. Our training data is based on QM information that is computed directly using density functional theory (DFT). To perform the fitting, we represent atomic environments using a set of rotationally, permutationally and reflection invariant parameters which act as the independent variables in our equations of non-parametric, non-linear regression. We develop a protocol for generating GAP models capable of describing lattice defects in metals by building a series of interatomic potentials for tungsten. We then demonstrate that a GAP potential based on a Smooth Overlap of Atomic Positions (SOAP) covariance function provides a description of the $\frac{1}{2} \langle 111 \rangle$ screw dislocation that is in agreement with the DFT model. We use this potential to simulate the mobility of $\frac{1}{2} \langle 111 \rangle$ screw dislocations by computing the Peierls barrier and model dislocation-vacancy interactions to QM accuracy in a system containing more than 100,000 atoms.

Self-pulsation in a two-section DFB laser with a varied ridge width

A 1.55 mu m InGaAsP-InP two-section DFB laser with a variable ridge width has been fabricated. Self-pulsations with frequencies around 3 GHz and 40 GHz are observed. The pulsation mechanisms related to the two frequencies are discussed and the tunability of generated self-pulsations is studied.

A ridge width varied two-section index-coupled DFB self-pulsation laser with a wide continuously tunable frequency range

A 1.55 mu m InGaAsP-InP index-coupled two-section DFB self-pulsation laser (SPL) with a varied ridge width has been fabricated. A record wide self-pulsation tuning range above 450 GHz has been achieved for this index-coupled DFB SPL. Furthermore, frequency locking to an optically injected modulated signal is successfully demonstrated.

1.55-mu m ridge DFB laser integrated with a buried-ridge-stripe dual-core spot-size converter by quantum-well intermixing

A ridge distributed feedback laser monolithically integrated with a buried-ridge-stripe spot-size converter operating at 1.55 mu m was successfully fabricated by means of low-energy ion implantation quantum-well intermixing and dual-core technologies. The passive waveguide was optically combined with a laterally exponentially tapered active core to control the mode size. The devices emit in a single transverse and single longitudinal mode with a sidemode suppression ratio of 38.0 dB. The threshold current was 25 mA. The beam divergence angles in the horizontal and vertical directions were as small as 8.0 degrees x 12.6 degrees, respectively, resulting in 3.0-dB coupling loss with a cleaved single-mode optical fiber.

Ridge waveguide laser integrated with spot-size converter in a single step epitaxial for high coupling efficiency to single-mode fibres

A 1.55-mum laser diode integrated with a spot-size converter was fabricated in a single step epitaxial by using the conventional photolithography and chemical wet etching process. The device was constructed by a conventional ridge waveguide active layer and a larger passive ridge-waveguide layer. The threshold current was 40 mA together with high slope efficiency of 0.24 W/A. The beam divergence angles in the horizontal and vertical directions were as small as 12.0degrees x 15.0degrees, respectively, resulting in about 3.2-dB coupling losses with a cleaved optical fibre.

Electroabsorption modulator integrated with buried-ridge-stripe dual-core spot-size converter by quantum-well intermixing

A single shallow ridge electroabsorption modulator monolithically integrated with a buried-ridge-stripe dual-core spot-size converter at the input and output port was fabricated by combining quantum-well intermixing and dual-core integration techniques simultaneously, using only a two-step low-pressure metal-organic vapor phase epitaxial process, conventional photolithography, and a chemical wet etching process. The optical insertion loss of the modulator in the on-state and the dc extinction ratio between 0 and -3 V at 1550 nm was -7.5 and 16 dB, respectively. The 3-dB modulation bandwidth was more than 10.0 GHz in electrical-optical response.

Thermal lensing effect in ridge structure InGaN multiple quantum well laser diodes

Time-resolved light-current curves, spectra, and far-field distributions of ridge structure InGaN multiple quantum well laser diodes grown on sapphire substrate are measured with a temporal resolution of 0.1 ns under a pulsed current condition. Results show that the thermal lensing effect clearly improves the confinement of the higher order modes. The thermal lens leads to a lower threshold current for the higher order modes, a higher slope efficiency, and a change in the lasing mode of the device. The threshold current for the higher modes decreases by about 5 mA in every 10 ns in a pulse, and the slope efficiency increases by 7.5 times on the average when higher modes lase. (c) 2006 American Institute of Physics.

1.55-mu m Ridge DFB laser and electroabsorption modulator integrated with buried-ridge-stripe dual-waveguide spot-size converter output

A 1.55-mu m ridge distributed feedback laser and electroabsorption modulator monolithically integrated with a buried-ridge-stripe dual-waveguide spot-size converter (SSC) at the output port for low-loss coupling to a cleaved single-mode optical fiber was fabricated by means of selective area growth, quantum-well intermixing, and dual-core technologies. These devices exhibit threshold current of 28 mA, 3-dB modulation bandwidth of 12.0 GHz, modulator extinction ratios of 25.0-dB dc. The output beam divergence angles of the SSC in the horizontal and vertical directions are as small as 8.0 degrees x 12.6 degrees, respectively, resulting in 3.2-dB coupling loss with a cleaved single-mode optical fiber.

Design considerations for a silicon-based p-i-n phase modulator in a double ridge waveguide with side isolating grooves

We present detail design considerations and simulation results of a forward biased carrier injection p-i-n modulator integrated on SOI rib waveguides. To minimize the free carrier absorption loss while keeping the comparatively small lateral dimensions of the modulator as required for high speed operation, we proposed two structural improvements, namely the double ridge (terrace ridge) structure and the isolating grooves at both sides of the double ridge. With improved carrier injection and optical confinement structure, the simulated modulator response time is in sub-ns range and absorption loss is minimized.