3D multi-car tracking based on monocular video

We propose a system that can reliably track multiple cars in congested traffic environments. Our system's key basis is the implementation of a sequential Monte Carlo algorithm, which introduces robustness against problems arising due to the proximity between vehicles. By directly modelling occlusions and collisions between cars we obtain promising results on an urban traffic dataset. Extensions to this initial framework are also suggested. © 2010 IEEE.

An Online Expectation-Maximization Algorithm for Changepoint Models

Changepoint models are widely used to model the heterogeneity of sequential data. We present a novel sequential Monte Carlo (SMC) online Expectation-Maximization (EM) algorithm for estimating the static parameters of such models. The SMC online EM algorithm has a cost per time which is linear in the number of particles and could be particularly important when the data is representable as a long sequence of observations, since it drastically reduces the computational requirements for implementation. We present an asymptotic analysis for the stability of the SMC estimates used in the online EM algorithm and demonstrate the performance of this scheme using both simulated and real data originating from DNA analysis.

An Online Expectation-Maximization Algorithm for Changepoint Models

Changepoint models are widely used to model the heterogeneity of sequential data. We present a novel sequential Monte Carlo (SMC) online Expectation-Maximization (EM) algorithm for estimating the static parameters of such models. The SMC online EM algorithm has a cost per time which is linear in the number of particles and could be particularly important when the data is representable as a long sequence of observations, since it drastically reduces the computational requirements for implementation. We present an asymptotic analysis for the stability of the SMC estimates used in the online EM algorithm and demonstrate the performance of this scheme using both simulated and real data originating from DNA analysis.

Comparison of stochastic methods for control in air traffic management

This paper provides a direct comparison of two stochastic optimisation techniques (Markov Chain Monte Carlo and Sequential Monte Carlo) when applied to the problem of conflict resolution and aircraft trajectory control in air traffic management. The two methods are then also compared to another existing technique of Mixed-Integer Linear Programming which is also popular in distributed control. © 2011 IFAC.

An online expectation-maximisation algorithm for nonnegative matrix factorisation models

In this paper we formulate the nonnegative matrix factorisation (NMF) problem as a maximum likelihood estimation problem for hidden Markov models and propose online expectation-maximisation (EM) algorithms to estimate the NMF and the other unknown static parameters. We also propose a sequential Monte Carlo approximation of our online EM algorithm. We show the performance of the proposed method with two numerical examples. © 2012 IFAC.

Approximate Bayesian Computation for Smoothing

We consider a method for approximate inference in hidden Markov models (HMMs). The method circumvents the need to evaluate conditional densities of observations given the hidden states. It may be considered an instance of Approximate Bayesian Computation (ABC) and it involves the introduction of auxiliary variables valued in the same space as the observations. The quality of the approximation may be controlled to arbitrary precision through a parameter ε > 0. We provide theoretical results which quantify, in terms of ε, the ABC error in approximation of expectations of additive functionals with respect to the smoothing distributions. Under regularity assumptions, this error is, where n is the number of time steps over which smoothing is performed. For numerical implementation, we adopt the forward-only sequential Monte Carlo (SMC) scheme of [14] and quantify the combined error from the ABC and SMC approximations. This forms some of the first quantitative results for ABC methods which jointly treat the ABC and simulation errors, with a finite number of data and simulated samples. © Taylor & Francis Group, LLC.

Likelihood-free estimation of model evidence

Statistical methods of inference typically require the likelihood function to be computable in a reasonable amount of time. The class of “likelihood-free” methods termed Approximate Bayesian Computation (ABC) is able to eliminate this requirement, replacing the evaluation of the likelihood with simulation from it. Likelihood-free methods have gained in efficiency and popularity in the past few years, following their integration with Markov Chain Monte Carlo (MCMC) and Sequential Monte Carlo (SMC) in order to better explore the parameter space. They have been applied primarily to estimating the parameters of a given model, but can also be used to compare models. Here we present novel likelihood-free approaches to model comparison, based upon the independent estimation of the evidence of each model under study. Key advantages of these approaches over previous techniques are that they allow the exploitation of MCMC or SMC algorithms for exploring the parameter space, and that they do not require a sampler able to mix between models. We validate the proposed methods using a simple exponential family problem before providing a realistic problem from human population genetics: the comparison of different demographic models based upon genetic data from the Y chromosome.

Analysis and compensation of power amplifier nonlinearity in MIMO transmit diversity systems

The nonlinearity of high-power amplifiers (HPAs) has a crucial effect on the performance of multiple-input-multiple-output (MIMO) systems. In this paper, we investigate the performance of MIMO orthogonal space-time block coding (OSTBC) systems in the presence of nonlinear HPAs. Specifically, we propose a constellation-based compensation method for HPA nonlinearity in the case with knowledge of the HPA parameters at the transmitter and receiver, where the constellation and decision regions of the distorted transmitted signal are derived in advance. Furthermore, in the scenario without knowledge of the HPA parameters, a sequential Monte Carlo (SMC)-based compensation method for the HPA nonlinearity is proposed, which first estimates the channel-gain matrix by means of the SMC method and then uses the SMC-based algorithm to detect the desired signal. The performance of the MIMO-OSTBC system under study is evaluated in terms of average symbol error probability (SEP), total degradation (TD) and system capacity, in uncorrelated Nakagami-m fading channels. Numerical and simulation results are provided and show the effects on performance of several system parameters, such as the parameters of the HPA model, output back-off (OBO) of nonlinear HPA, numbers of transmit and receive antennas, modulation order of quadrature amplitude modulation (QAM), and number of SMC samples. In particular, it is shown that the constellation-based compensation method can efficiently mitigate the effect of HPA nonlinearity with low complexity and that the SMC-based detection scheme is efficient to compensate for HPA nonlinearity in the case without knowledge of the HPA parameters.

A nonparametric ensemble transform method for Bayesian inference

Many applications, such as intermittent data assimilation, lead to a recursive application of Bayesian inference within a Monte Carlo context. Popular data assimilation algorithms include sequential Monte Carlo methods and ensemble Kalman filters (EnKFs). These methods differ in the way Bayesian inference is implemented. Sequential Monte Carlo methods rely on importance sampling combined with a resampling step, while EnKFs utilize a linear transformation of Monte Carlo samples based on the classic Kalman filter. While EnKFs have proven to be quite robust even for small ensemble sizes, they are not consistent since their derivation relies on a linear regression ansatz. In this paper, we propose another transform method, which does not rely on any a priori assumptions on the underlying prior and posterior distributions. The new method is based on solving an optimal transportation problem for discrete random variables. © 2013, Society for Industrial and Applied Mathematics

Bayesian model comparison with un-normalised likelihoods

Models for which the likelihood function can be evaluated only up to a parameter-dependent unknown normalizing constant, such as Markov random field models, are used widely in computer science, statistical physics, spatial statistics, and network analysis. However, Bayesian analysis of these models using standard Monte Carlo methods is not possible due to the intractability of their likelihood functions. Several methods that permit exact, or close to exact, simulation from the posterior distribution have recently been developed. However, estimating the evidence and Bayes’ factors for these models remains challenging in general. This paper describes new random weight importance sampling and sequential Monte Carlo methods for estimating BFs that use simulation to circumvent the evaluation of the intractable likelihood, and compares them to existing methods. In some cases we observe an advantage in the use of biased weight estimates. An initial investigation into the theoretical and empirical properties of this class of methods is presented. Some support for the use of biased estimates is presented, but we advocate caution in the use of such estimates.

Continuum, Discrete, and Explicit Solvation Models for Describing the Low-Lying Absorption Spectrum of the Pterin Acid in Aqueous Environment

The absorption spectrum of the acid form of pterin in water was investigated theoretically. Different procedures using continuum, discrete, and explicit models were used to include the solvation effect on the absorption spectrum, characterized by two bands. The discrete and explicit models used Monte Carlo simulation to generate the liquid structure and time-dependent density functional theory (B3LYP/6-31G+(d)) to obtain the excitation energies. The discrete model failed to give the correct qualitative effect on the second absorption band. The continuum model, in turn, has given a correct qualitative picture and a semiquantitative description. The explicit use of 29 solvent molecules, forming a hydration shell of 6 angstrom, embedded in the electrostatic field of the remaining solvent molecules, gives absorption transitions at 3.67 and 4.59 eV in excellent agreement with the S(0)-S(1) and S(0)-S(2) absorption bands at of 3.66 and 4.59 eV, respectively, that characterize the experimental spectrum of pterin in water environment. (C) 2010 Wiley Periodicals, Inc. Int J Quantum Chem 110: 2371-2377, 2010

Electronic spectroscopy of biomolecules in solution: fluorescein dianion in water

A combined and sequential Monte Carlo-quantum mechanics methodology is used to describe the electronic absorption spectrum of the fluorescein dianion in water. Different sets of 100 statistically relevant configurations composed of the solute and several solvent molecules are sampled from the Monte Carlo simulation for a posteriori quantum mechanical calculations of the spectra. In the largest case the configurations are composed of fluorescein and 90 explicit water molecules embedded in the electrostatic field of all remaining water molecules within a distance of 11.3 angstrom. These configurations include 305 atoms and 842 valence electrons, justifying the use of a semi-empirical approach. The electronic spectrum is then calculated using the INDO/CIS method. The solvatochromic shift of fluorescein in water, compared with in isolation, is calculated using the discrete and explicit solvent models. The use of electrostatically embedded explicit water molecules, in INDO/CIS calculations, gives a good description of the spectral shift of the fluorescein dianion in aqueous environment. The results are verified to converge both statistically and with respect to the number of explicit solvent molecules used.

Polarization and Spectral Shift of Benzophenone in Supercritical Water

Monte Carlo simulation and quantum mechanics calculations based on the INDO/CIS and TD-DFT methods were utilized to study the solvatochromic shift of benzophenone when changing the environment from normal water to supercritical (P = 340.2 atm and T = 673 K) condition. Solute polarization increases the dipole moment of benzophenone, compared to gas phase, by 88 and 35% in normal and supercritical conditions, giving the in-solvent dipole value of 5.8 and 4.2 D, respectively. The average number of solute-solvent hydrogen bonds was analyzed, and a large decrease of 2.3 in normal water to only 0.8 in the supercritical environment was found. By using these polarized models of benzophenone in the two different conditions of water, we performed MC simulations to generate statistically uncorrelated configurations of the solute surrounded by the solvent molecules and subsequent quantum mechanics calculations on these configurations. When changing from normal to supercritical water environment, INDO/CIS calculations explicitly considering all valence electrons of the 235 solvent water molecules resulted in a solvatochromic shift of 1425 cm(-1) for the most intense transition of benzophenone, that is, slightly underestimated in comparison with the experimentally inferred result of 1700 cm(-1). TD-B3LYP/6-311+G(2d,p) calculations on the same configurations but with benzophenone electrostatically embedded in the 320 water molecules resulted in a solvatochromic shift of 1715 cm(-1) for this transition, in very good agreement with the experimental result. When using the unpolarized model of the benzophenone, this calculated solvatochromic shift was only 640 cm(-1). Additional calculations were also made by using BHandHLYP/6-311+G(2d,p) to analyze the effect of the asymptotic decay of the exchange functional. This study indicates that, contrary to the general expectation, there is a sizable solute polarization even in the low-density regime of supercritical condition and that the inclusion of this polarization is important for a reliable description of the spectral shifts considered here.

Polarization and solvatochromic shift of ortho-betaine in water

Betaine dyes are known to show very large transition energy shifts in different solvents. The ortho-betaine molecule - a simple two-ring prototype of the E-T(30) Reichardt dye - has been investigated theoretically from a combined statistical and quantum mechanics approach. Using sequential Monte Carlo (MC) simulations and MP2/cc-pVDZ calculations the in-water dipole moment of ortho-betaine is obtained as 12.30 +/- 0.05 D. This result shows a considerable increase of 75% compared to the in-vacuum dipole moment. For comparison, the use of a polarizable continuum model using the same MP2/cc-pVDZ leads to an in-water dipole moment of 11.6 D, in good agreement. This large polarization is incorporated in the classical potential for another MC simulation to generate solute-solvent configurations and to obtain the contribution of the polarization effect in the solvatochromic shift. Using statistically uncorrelated configurations and supermolecular INDO/CIS calculations, including the solute and, explicitly, 230 solvent water molecules, the statistically converged calculated shift is obtained here as 6360 cm(-1), in good agreement with the experimental result of 7550 cm(-1). (c) 2007 Elsevier B.V. All rights reserved.

Solvatochromic Shift of the pi -> pi* Transition in All-Trans, Cis-13, Cis-11, Cis-9, and Cis-7 Retinal Isomers Induced by Water and Methanol

The solvatochromic shift of the lowest singlet it pi -> pi* electronic transition in the all-trans, cis-13, cis-11, cis-9, and cis-7 retinal isomers were computed under the influence of water, methanol, and benzene solvents. Excitation energies were calculated in gas phase and in solution. The calculations in solution were performed considering the sequential Monte Carlo (MC) /Quantum Mechanical approach. The MC simulations were performed considering the full retinal isomer molecules and 900 water molecules, 900 methanol, or 400 benzene ones. The OPLS/AA parametrization was chosen for retinal, methanol, and benzene molecules and the SPC model was used for water one. From the MC calculations 100 independent configurations were selected, with 100 solvent molecules in thermodynamical equilibrium at T = 298.15 K. Average point-charges were obtained from those independent configurations for water, methanol, and benzene solvent. TDDFT and CASSCF//CASPT2 methodologies were used to compute the vertical excitation energy of the retinal isomers in different environment. (C) 2010 Wiley Periodicals, Inc. Int J Quantum Chem 110: 2076-2087, 2010