-
Towards Photographic Image Manipulation with Balanced Growing of Generative Autoencoders
Authors:
Ari Heljakka,
Arno Solin,
Juho Kannala
Abstract:
We present a generative autoencoder that provides fast encoding, faithful reconstructions (eg. retaining the identity of a face), sharp generated/reconstructed samples in high resolutions, and a well-structured latent space that supports semantic manipulation of the inputs. There are no current autoencoder or GAN models that satisfactorily achieve all of these. We build on the progressively growin…
▽ More
We present a generative autoencoder that provides fast encoding, faithful reconstructions (eg. retaining the identity of a face), sharp generated/reconstructed samples in high resolutions, and a well-structured latent space that supports semantic manipulation of the inputs. There are no current autoencoder or GAN models that satisfactorily achieve all of these. We build on the progressively growing autoencoder model PIONEER, for which we completely alter the training dynamics based on a careful analysis of recently introduced normalization schemes. We show significantly improved visual and quantitative results for face identity conservation in CelebAHQ. Our model achieves state-of-the-art disentanglement of latent space, both quantitatively and via realistic image attribute manipulations. On the LSUN Bedrooms dataset, we improve the disentanglement performance of the vanilla PIONEER, despite having a simpler model. Overall, our results indicate that the PIONEER networks provide a way towards photorealistic face manipulation.
△ Less
Submitted 20 February, 2020; v1 submitted 12 April, 2019;
originally announced April 2019.
-
Know Your Boundaries: Constraining Gaussian Processes by Variational Harmonic Features
Authors:
Arno Solin,
Manon Kok
Abstract:
Gaussian processes (GPs) provide a powerful framework for extrapolation, interpolation, and noise removal in regression and classification. This paper considers constraining GPs to arbitrarily-shaped domains with boundary conditions. We solve a Fourier-like generalised harmonic feature representation of the GP prior in the domain of interest, which both constrains the GP and attains a low-rank rep…
▽ More
Gaussian processes (GPs) provide a powerful framework for extrapolation, interpolation, and noise removal in regression and classification. This paper considers constraining GPs to arbitrarily-shaped domains with boundary conditions. We solve a Fourier-like generalised harmonic feature representation of the GP prior in the domain of interest, which both constrains the GP and attains a low-rank representation that is used for speeding up inference. The method scales as $\mathcal{O}(nm^2)$ in prediction and $\mathcal{O}(m^3)$ in hyperparameter learning for regression, where $n$ is the number of data points and $m$ the number of features. Furthermore, we make use of the variational approach to allow the method to deal with non-Gaussian likelihoods. The experiments cover both simulated and empirical data in which the boundary conditions allow for inclusion of additional physical information.
△ Less
Submitted 10 April, 2019;
originally announced April 2019.
-
Interpolation Consistency Training for Semi-Supervised Learning
Authors:
Vikas Verma,
Kenji Kawaguchi,
Alex Lamb,
Juho Kannala,
Arno Solin,
Yoshua Bengio,
David Lopez-Paz
Abstract:
We introduce Interpolation Consistency Training (ICT), a simple and computation efficient algorithm for training Deep Neural Networks in the semi-supervised learning paradigm. ICT encourages the prediction at an interpolation of unlabeled points to be consistent with the interpolation of the predictions at those points. In classification problems, ICT moves the decision boundary to low-density reg…
▽ More
We introduce Interpolation Consistency Training (ICT), a simple and computation efficient algorithm for training Deep Neural Networks in the semi-supervised learning paradigm. ICT encourages the prediction at an interpolation of unlabeled points to be consistent with the interpolation of the predictions at those points. In classification problems, ICT moves the decision boundary to low-density regions of the data distribution. Our experiments show that ICT achieves state-of-the-art performance when applied to standard neural network architectures on the CIFAR-10 and SVHN benchmark datasets. Our theoretical analysis shows that ICT corresponds to a certain type of data-adaptive regularization with unlabeled points which reduces overfitting to labeled points under high confidence values.
△ Less
Submitted 19 October, 2022; v1 submitted 9 March, 2019;
originally announced March 2019.
-
Unstructured Multi-View Depth Estimation Using Mask-Based Multiplane Representation
Authors:
Yuxin Hou,
Arno Solin,
Juho Kannala
Abstract:
This paper presents a novel method, MaskMVS, to solve depth estimation for unstructured multi-view image-pose pairs. In the plane-sweep procedure, the depth planes are sampled by histogram matching that ensures covering the depth range of interest. Unlike other plane-sweep methods, we do not rely on a cost metric to explicitly build the cost volume, but instead infer a multiplane mask representati…
▽ More
This paper presents a novel method, MaskMVS, to solve depth estimation for unstructured multi-view image-pose pairs. In the plane-sweep procedure, the depth planes are sampled by histogram matching that ensures covering the depth range of interest. Unlike other plane-sweep methods, we do not rely on a cost metric to explicitly build the cost volume, but instead infer a multiplane mask representation which regularizes the learning. Compared to many previous approaches, we show that our method is lightweight and generalizes well without requiring excessive training. We outperform the current state-of-the-art and show results on the sun3d, scenes11, MVS, and RGBD test data sets.
△ Less
Submitted 10 April, 2019; v1 submitted 6 February, 2019;
originally announced February 2019.
-
End-to-End Probabilistic Inference for Nonstationary Audio Analysis
Authors:
William J. Wilkinson,
Michael Riis Andersen,
Joshua D. Reiss,
Dan Stowell,
Arno Solin
Abstract:
A typical audio signal processing pipeline includes multiple disjoint analysis stages, including calculation of a time-frequency representation followed by spectrogram-based feature analysis. We show how time-frequency analysis and nonnegative matrix factorisation can be jointly formulated as a spectral mixture Gaussian process model with nonstationary priors over the amplitude variance parameters…
▽ More
A typical audio signal processing pipeline includes multiple disjoint analysis stages, including calculation of a time-frequency representation followed by spectrogram-based feature analysis. We show how time-frequency analysis and nonnegative matrix factorisation can be jointly formulated as a spectral mixture Gaussian process model with nonstationary priors over the amplitude variance parameters. Further, we formulate this nonlinear model's state space representation, making it amenable to infinite-horizon Gaussian process regression with approximate inference via expectation propagation, which scales linearly in the number of time steps and quadratically in the state dimensionality. By doing so, we are able to process audio signals with hundreds of thousands of data points. We demonstrate, on various tasks with empirical data, how this inference scheme outperforms more standard techniques that rely on extended Kalman filtering.
△ Less
Submitted 27 April, 2019; v1 submitted 31 January, 2019;
originally announced January 2019.
-
Infinite-Horizon Gaussian Processes
Authors:
Arno Solin,
James Hensman,
Richard E. Turner
Abstract:
Gaussian processes provide a flexible framework for forecasting, removing noise, and interpreting long temporal datasets. State space modelling (Kalman filtering) enables these non-parametric models to be deployed on long datasets by reducing the complexity to linear in the number of data points. The complexity is still cubic in the state dimension $m$ which is an impediment to practical applicati…
▽ More
Gaussian processes provide a flexible framework for forecasting, removing noise, and interpreting long temporal datasets. State space modelling (Kalman filtering) enables these non-parametric models to be deployed on long datasets by reducing the complexity to linear in the number of data points. The complexity is still cubic in the state dimension $m$ which is an impediment to practical application. In certain special cases (Gaussian likelihood, regular spacing) the GP posterior will reach a steady posterior state when the data are very long. We leverage this and formulate an inference scheme for GPs with general likelihoods, where inference is based on single-sweep EP (assumed density filtering). The infinite-horizon model tackles the cubic cost in the state dimensionality and reduces the cost in the state dimension $m$ to $\mathcal{O}(m^2)$ per data point. The model is extended to online-learning of hyperparameters. We show examples for large finite-length modelling problems, and present how the method runs in real-time on a smartphone on a continuous data stream updated at 100~Hz.
△ Less
Submitted 15 November, 2018;
originally announced November 2018.
-
Unifying Probabilistic Models for Time-Frequency Analysis
Authors:
William J. Wilkinson,
Michael Riis Andersen,
Joshua D. Reiss,
Dan Stowell,
Arno Solin
Abstract:
In audio signal processing, probabilistic time-frequency models have many benefits over their non-probabilistic counterparts. They adapt to the incoming signal, quantify uncertainty, and measure correlation between the signal's amplitude and phase information, making time domain resynthesis straightforward. However, these models are still not widely used since they come at a high computational cos…
▽ More
In audio signal processing, probabilistic time-frequency models have many benefits over their non-probabilistic counterparts. They adapt to the incoming signal, quantify uncertainty, and measure correlation between the signal's amplitude and phase information, making time domain resynthesis straightforward. However, these models are still not widely used since they come at a high computational cost, and because they are formulated in such a way that it can be difficult to interpret all the modelling assumptions. By showing their equivalence to Spectral Mixture Gaussian processes, we illuminate the underlying model assumptions and provide a general framework for constructing more complex models that better approximate real-world signals. Our interpretation makes it intuitive to inspect, compare, and alter the models since all prior knowledge is encoded in the Gaussian process kernel functions. We utilise a state space representation to perform efficient inference via Kalman smoothing, and we demonstrate how our interpretation allows for efficient parameter learning in the frequency domain.
△ Less
Submitted 12 February, 2019; v1 submitted 6 November, 2018;
originally announced November 2018.
-
Deep Learning Based Speed Estimation for Constraining Strapdown Inertial Navigation on Smartphones
Authors:
Santiago Cortés,
Arno Solin,
Juho Kannala
Abstract:
Strapdown inertial navigation systems are sensitive to the quality of the data provided by the accelerometer and gyroscope. Low-grade IMUs in handheld smart-devices pose a problem for inertial odometry on these devices. We propose a scheme for constraining the inertial odometry problem by complementing non-linear state estimation by a CNN-based deep-learning model for inferring the momentary speed…
▽ More
Strapdown inertial navigation systems are sensitive to the quality of the data provided by the accelerometer and gyroscope. Low-grade IMUs in handheld smart-devices pose a problem for inertial odometry on these devices. We propose a scheme for constraining the inertial odometry problem by complementing non-linear state estimation by a CNN-based deep-learning model for inferring the momentary speed based on a window of IMU samples. We show the feasibility of the model using a wide range of data from an iPhone, and present proof-of-concept results for how the model can be combined with an inertial navigation system for three-dimensional inertial navigation.
△ Less
Submitted 10 August, 2018;
originally announced August 2018.
-
ADVIO: An authentic dataset for visual-inertial odometry
Authors:
Santiago Cortés,
Arno Solin,
Esa Rahtu,
Juho Kannala
Abstract:
The lack of realistic and open benchmarking datasets for pedestrian visual-inertial odometry has made it hard to pinpoint differences in published methods. Existing datasets either lack a full six degree-of-freedom ground-truth or are limited to small spaces with optical tracking systems. We take advantage of advances in pure inertial navigation, and develop a set of versatile and challenging real…
▽ More
The lack of realistic and open benchmarking datasets for pedestrian visual-inertial odometry has made it hard to pinpoint differences in published methods. Existing datasets either lack a full six degree-of-freedom ground-truth or are limited to small spaces with optical tracking systems. We take advantage of advances in pure inertial navigation, and develop a set of versatile and challenging real-world computer vision benchmark sets for visual-inertial odometry. For this purpose, we have built a test rig equipped with an iPhone, a Google Pixel Android phone, and a Google Tango device. We provide a wide range of raw sensor data that is accessible on almost any modern-day smartphone together with a high-quality ground-truth track. We also compare resulting visual-inertial tracks from Google Tango, ARCore, and Apple ARKit with two recent methods published in academic forums. The data sets cover both indoor and outdoor cases, with stairs, escalators, elevators, office environments, a shop** mall, and metro station.
△ Less
Submitted 25 July, 2018;
originally announced July 2018.
-
Pioneer Networks: Progressively Growing Generative Autoencoder
Authors:
Ari Heljakka,
Arno Solin,
Juho Kannala
Abstract:
We introduce a novel generative autoencoder network model that learns to encode and reconstruct images with high quality and resolution, and supports smooth random sampling from the latent space of the encoder. Generative adversarial networks (GANs) are known for their ability to simulate random high-quality images, but they cannot reconstruct existing images. Previous works have attempted to exte…
▽ More
We introduce a novel generative autoencoder network model that learns to encode and reconstruct images with high quality and resolution, and supports smooth random sampling from the latent space of the encoder. Generative adversarial networks (GANs) are known for their ability to simulate random high-quality images, but they cannot reconstruct existing images. Previous works have attempted to extend GANs to support such inference but, so far, have not delivered satisfactory high-quality results. Instead, we propose the Progressively Growing Generative Autoencoder (PIONEER) network which achieves high-quality reconstruction with $128{\times}128$ images without requiring a GAN discriminator. We merge recent techniques for progressively building up the parts of the network with the recently introduced adversarial encoder-generator network. The ability to reconstruct input images is crucial in many real-world applications, and allows for precise intelligent manipulation of existing images. We show promising results in image synthesis and inference, with state-of-the-art results in CelebA inference tasks.
△ Less
Submitted 9 October, 2018; v1 submitted 9 July, 2018;
originally announced July 2018.
-
Robust Gyroscope-Aided Camera Self-Calibration
Authors:
Santiago Cortés Reina,
Arno Solin,
Juho Kannala
Abstract:
Camera calibration for estimating the intrinsic parameters and lens distortion is a prerequisite for various monocular vision applications including feature tracking and video stabilization. This application paper proposes a model for estimating the parameters on the fly by fusing gyroscope and camera data, both readily available in modern day smartphones. The model is based on joint estimation of…
▽ More
Camera calibration for estimating the intrinsic parameters and lens distortion is a prerequisite for various monocular vision applications including feature tracking and video stabilization. This application paper proposes a model for estimating the parameters on the fly by fusing gyroscope and camera data, both readily available in modern day smartphones. The model is based on joint estimation of visual feature positions, camera parameters, and the camera pose, the movement of which is assumed to follow the movement predicted by the gyroscope. Our model assumes the camera movement to be free, but continuous and differentiable, and individual features are assumed to stay stationary. The estimation is performed online using an extended Kalman filter, and it is shown to outperform existing methods in robustness and insensitivity to initialization. We demonstrate the method using simulated data and empirical data from an iPad.
△ Less
Submitted 31 May, 2018;
originally announced May 2018.
-
Scalable Magnetic Field SLAM in 3D Using Gaussian Process Maps
Authors:
Manon Kok,
Arno Solin
Abstract:
We present a method for scalable and fully 3D magnetic field simultaneous localisation and map** (SLAM) using local anomalies in the magnetic field as a source of position information. These anomalies are due to the presence of ferromagnetic material in the structure of buildings and in objects such as furniture. We represent the magnetic field map using a Gaussian process model and take well-kn…
▽ More
We present a method for scalable and fully 3D magnetic field simultaneous localisation and map** (SLAM) using local anomalies in the magnetic field as a source of position information. These anomalies are due to the presence of ferromagnetic material in the structure of buildings and in objects such as furniture. We represent the magnetic field map using a Gaussian process model and take well-known physical properties of the magnetic field into account. We build local maps using three-dimensional hexagonal block tiling. To make our approach computationally tractable we use reduced-rank Gaussian process regression in combination with a Rao-Blackwellised particle filter. We show that it is possible to obtain accurate position and orientation estimates using measurements from a smartphone, and that our approach provides a scalable magnetic field SLAM algorithm in terms of both computational complexity and map storage.
△ Less
Submitted 10 June, 2018; v1 submitted 5 April, 2018;
originally announced April 2018.
-
Computationally Inferred Genealogical Networks Uncover Long-Term Trends in Assortative Mating
Authors:
Eric Malmi,
Aristides Gionis,
Arno Solin
Abstract:
Genealogical networks, also known as family trees or population pedigrees, are commonly studied by genealogists wanting to know about their ancestry, but they also provide a valuable resource for disciplines such as digital demography, genetics, and computational social science. These networks are typically constructed by hand through a very time-consuming process, which requires comparing large n…
▽ More
Genealogical networks, also known as family trees or population pedigrees, are commonly studied by genealogists wanting to know about their ancestry, but they also provide a valuable resource for disciplines such as digital demography, genetics, and computational social science. These networks are typically constructed by hand through a very time-consuming process, which requires comparing large numbers of historical records manually. We develop computational methods for automatically inferring large-scale genealogical networks. A comparison with human-constructed networks attests to the accuracy of the proposed methods. To demonstrate the applicability of the inferred large-scale genealogical networks, we present a longitudinal analysis on the mating patterns observed in a network. This analysis shows a consistent tendency of people choosing a spouse with a similar socioeconomic status, a phenomenon known as assortative mating. Interestingly, we do not observe this tendency to consistently decrease (nor increase) over our study period of 150 years.
△ Less
Submitted 16 February, 2018;
originally announced February 2018.
-
Recursive Chaining of Reversible Image-to-image Translators For Face Aging
Authors:
Ari Heljakka,
Arno Solin,
Juho Kannala
Abstract:
This paper addresses the modeling and simulation of progressive changes over time, such as human face aging. By treating the age phases as a sequence of image domains, we construct a chain of transformers that map images from one age domain to the next. Leveraging recent adversarial image translation methods, our approach requires no training samples of the same individual at different ages. Here,…
▽ More
This paper addresses the modeling and simulation of progressive changes over time, such as human face aging. By treating the age phases as a sequence of image domains, we construct a chain of transformers that map images from one age domain to the next. Leveraging recent adversarial image translation methods, our approach requires no training samples of the same individual at different ages. Here, the model must be flexible enough to translate a child face to a young adult, and all the way through the adulthood to old age. We find that some transformers in the chain can be recursively applied on their own output to cover multiple phases, compressing the chain. The structure of the chain also unearths information about the underlying physical process. We demonstrate the performance of our method with precise and intuitive metrics, and visually match with the face aging state-of-the-art.
△ Less
Submitted 6 August, 2018; v1 submitted 14 February, 2018;
originally announced February 2018.
-
State Space Gaussian Processes with Non-Gaussian Likelihood
Authors:
Hannes Nickisch,
Arno Solin,
Alexander Grigorievskiy
Abstract:
We provide a comprehensive overview and tooling for GP modeling with non-Gaussian likelihoods using state space methods. The state space formulation allows for solving one-dimensional GP models in $\mathcal{O}(n)$ time and memory complexity. While existing literature has focused on the connection between GP regression and state space methods, the computational primitives allowing for inference usi…
▽ More
We provide a comprehensive overview and tooling for GP modeling with non-Gaussian likelihoods using state space methods. The state space formulation allows for solving one-dimensional GP models in $\mathcal{O}(n)$ time and memory complexity. While existing literature has focused on the connection between GP regression and state space methods, the computational primitives allowing for inference using general likelihoods in combination with the Laplace approximation (LA), variational Bayes (VB), and assumed density filtering (ADF, a.k.a. single-sweep expectation propagation, EP) schemes has been largely overlooked. We present means of combining the efficient $\mathcal{O}(n)$ state space methodology with existing inference methods. We extend existing methods, and provide unifying code implementing all approaches.
△ Less
Submitted 5 July, 2018; v1 submitted 13 February, 2018;
originally announced February 2018.
-
PIVO: Probabilistic Inertial-Visual Odometry for Occlusion-Robust Navigation
Authors:
Arno Solin,
Santiago Cortes,
Esa Rahtu,
Juho Kannala
Abstract:
This paper presents a novel method for visual-inertial odometry. The method is based on an information fusion framework employing low-cost IMU sensors and the monocular camera in a standard smartphone. We formulate a sequential inference scheme, where the IMU drives the dynamical model and the camera frames are used in coupling trailing sequences of augmented poses. The novelty in the model is in…
▽ More
This paper presents a novel method for visual-inertial odometry. The method is based on an information fusion framework employing low-cost IMU sensors and the monocular camera in a standard smartphone. We formulate a sequential inference scheme, where the IMU drives the dynamical model and the camera frames are used in coupling trailing sequences of augmented poses. The novelty in the model is in taking into account all the cross-terms in the updates, thus propagating the inter-connected uncertainties throughout the model. Stronger coupling between the inertial and visual data sources leads to robustness against occlusion and feature-poor environments. We demonstrate results on data collected with an iPhone and provide comparisons against the Tango device and using the EuRoC data set.
△ Less
Submitted 23 January, 2018; v1 submitted 2 August, 2017;
originally announced August 2017.
-
Inertial Odometry on Handheld Smartphones
Authors:
Arno Solin,
Santiago Cortes,
Esa Rahtu,
Juho Kannala
Abstract:
Building a complete inertial navigation system using the limited quality data provided by current smartphones has been regarded challenging, if not impossible. This paper shows that by careful crafting and accounting for the weak information in the sensor samples, smartphones are capable of pure inertial navigation. We present a probabilistic approach for orientation and use-case free inertial odo…
▽ More
Building a complete inertial navigation system using the limited quality data provided by current smartphones has been regarded challenging, if not impossible. This paper shows that by careful crafting and accounting for the weak information in the sensor samples, smartphones are capable of pure inertial navigation. We present a probabilistic approach for orientation and use-case free inertial odometry, which is based on double-integrating rotated accelerations. The strength of the model is in learning additive and multiplicative IMU biases online. We are able to track the phone position, velocity, and pose in real-time and in a computationally lightweight fashion by solving the inference with an extended Kalman filter. The information fusion is completed with zero-velocity updates (if the phone remains stationary), altitude correction from barometric pressure readings (if available), and pseudo-updates constraining the momentary speed. We demonstrate our approach using an iPad and iPhone in several indoor dead-reckoning applications and in a measurement tool setup.
△ Less
Submitted 7 June, 2018; v1 submitted 1 March, 2017;
originally announced March 2017.
-
Variational Fourier features for Gaussian processes
Authors:
James Hensman,
Nicolas Durrande,
Arno Solin
Abstract:
This work brings together two powerful concepts in Gaussian processes: the variational approach to sparse approximation and the spectral representation of Gaussian processes. This gives rise to an approximation that inherits the benefits of the variational approach but with the representational power and computational scalability of spectral representations. The work hinges on a key result that th…
▽ More
This work brings together two powerful concepts in Gaussian processes: the variational approach to sparse approximation and the spectral representation of Gaussian processes. This gives rise to an approximation that inherits the benefits of the variational approach but with the representational power and computational scalability of spectral representations. The work hinges on a key result that there exist spectral features related to a finite domain of the Gaussian process which exhibit almost-independent covariances. We derive these expressions for Matern kernels in one dimension, and generalize to more dimensions using kernels with specific structures. Under the assumption of additive Gaussian noise, our method requires only a single pass through the dataset, making for very fast and accurate computation. We fit a model to 4 million training points in just a few minutes on a standard laptop. With non-conjugate likelihoods, our MCMC scheme reduces the cost of computation from O(NM2) (for a sparse Gaussian process) to O(NM) per iteration, where N is the number of data and M is the number of features.
△ Less
Submitted 8 November, 2017; v1 submitted 21 November, 2016;
originally announced November 2016.
-
Regularizing Solutions to the MEG Inverse Problem Using Space-Time Separable Covariance Functions
Authors:
Arno Solin,
Pasi Jylänki,
Jaakko Kauramäki,
Tom Heskes,
Marcel A. J. van Gerven,
Simo Särkkä
Abstract:
In magnetoencephalography (MEG) the conventional approach to source reconstruction is to solve the underdetermined inverse problem independently over time and space. Here we present how the conventional approach can be extended by regularizing the solution in space and time by a Gaussian process (Gaussian random field) model. Assuming a separable covariance function in space and time, the computat…
▽ More
In magnetoencephalography (MEG) the conventional approach to source reconstruction is to solve the underdetermined inverse problem independently over time and space. Here we present how the conventional approach can be extended by regularizing the solution in space and time by a Gaussian process (Gaussian random field) model. Assuming a separable covariance function in space and time, the computational complexity of the proposed model becomes (without any further assumptions or restrictions) $\mathcal{O}(t^3 + n^3 + m^2n)$, where $t$ is the number of time steps, $m$ is the number of sources, and $n$ is the number of sensors. We apply the method to both simulated and empirical data, and demonstrate the efficiency and generality of our Bayesian source reconstruction approach which subsumes various classical approaches in the literature.
△ Less
Submitted 17 April, 2016;
originally announced April 2016.
-
Nonlinear State Space Model Identification Using a Regularized Basis Function Expansion
Authors:
Andreas Svensson,
Thomas B. Schön,
Arno Solin,
Simo Särkkä
Abstract:
This paper is concerned with black-box identification of nonlinear state space models. By using a basis function expansion within the state space model, we obtain a flexible structure. The model is identified using an expectation maximization approach, where the states and the parameters are updated iteratively in such a way that a maximum likelihood estimate is obtained. We use recent particle me…
▽ More
This paper is concerned with black-box identification of nonlinear state space models. By using a basis function expansion within the state space model, we obtain a flexible structure. The model is identified using an expectation maximization approach, where the states and the parameters are updated iteratively in such a way that a maximum likelihood estimate is obtained. We use recent particle methods with sound theoretical properties to infer the states, whereas the model parameters can be updated using closed-form expressions by exploiting the fact that our model is linear in the parameters. Not to over-fit the flexible model to the data, we also propose a regularization scheme without increasing the computational burden. Importantly, this opens up for systematic use of regularization in nonlinear state space models. We conclude by evaluating our proposed approach on one simulation example and two real-data problems.
△ Less
Submitted 2 October, 2015;
originally announced October 2015.
-
Modeling and interpolation of the ambient magnetic field by Gaussian processes
Authors:
Arno Solin,
Manon Kok,
Niklas Wahlström,
Thomas B. Schön,
Simo Särkkä
Abstract:
Anomalies in the ambient magnetic field can be used as features in indoor positioning and navigation. By using Maxwell's equations, we derive and present a Bayesian non-parametric probabilistic modeling approach for interpolation and extrapolation of the magnetic field. We model the magnetic field components jointly by imposing a Gaussian process (GP) prior on the latent scalar potential of the ma…
▽ More
Anomalies in the ambient magnetic field can be used as features in indoor positioning and navigation. By using Maxwell's equations, we derive and present a Bayesian non-parametric probabilistic modeling approach for interpolation and extrapolation of the magnetic field. We model the magnetic field components jointly by imposing a Gaussian process (GP) prior on the latent scalar potential of the magnetic field. By rewriting the GP model in terms of a Hilbert space representation, we circumvent the computational pitfalls associated with GP modeling and provide a computationally efficient and physically justified modeling tool for the ambient magnetic field. The model allows for sequential updating of the estimate and time-dependent changes in the magnetic field. The model is shown to work well in practice in different applications: we demonstrate map** of the magnetic field both with an inexpensive Raspberry Pi powered robot and on foot using a standard smartphone.
△ Less
Submitted 21 March, 2018; v1 submitted 15 September, 2015;
originally announced September 2015.
-
Computationally Efficient Bayesian Learning of Gaussian Process State Space Models
Authors:
Andreas Svensson,
Arno Solin,
Simo Särkkä,
Thomas B. Schön
Abstract:
Gaussian processes allow for flexible specification of prior assumptions of unknown dynamics in state space models. We present a procedure for efficient Bayesian learning in Gaussian process state space models, where the representation is formed by projecting the problem onto a set of approximate eigenfunctions derived from the prior covariance structure. Learning under this family of models can b…
▽ More
Gaussian processes allow for flexible specification of prior assumptions of unknown dynamics in state space models. We present a procedure for efficient Bayesian learning in Gaussian process state space models, where the representation is formed by projecting the problem onto a set of approximate eigenfunctions derived from the prior covariance structure. Learning under this family of models can be conducted using a carefully crafted particle MCMC algorithm. This scheme is computationally efficient and yet allows for a fully Bayesian treatment of the problem. Compared to conventional system identification tools or existing learning methods, we show competitive performance and reliable quantification of uncertainties in the model.
△ Less
Submitted 15 April, 2016; v1 submitted 7 June, 2015;
originally announced June 2015.
-
Sigma-Point Filtering and Smoothing Based Parameter Estimation in Nonlinear Dynamic Systems
Authors:
Juho Kokkala,
Arno Solin,
Simo Särkkä
Abstract:
We consider approximate maximum likelihood parameter estimation in nonlinear state-space models. We discuss both direct optimization of the likelihood and expectation--maximization (EM). For EM, we also give closed-form expressions for the maximization step in a class of models that are linear in parameters and have additive noise. To obtain approximations to the filtering and smoothing distributi…
▽ More
We consider approximate maximum likelihood parameter estimation in nonlinear state-space models. We discuss both direct optimization of the likelihood and expectation--maximization (EM). For EM, we also give closed-form expressions for the maximization step in a class of models that are linear in parameters and have additive noise. To obtain approximations to the filtering and smoothing distributions needed in the likelihood-maximization methods, we focus on using Gaussian filtering and smoothing algorithms that employ sigma-points to approximate the required integrals. We discuss different sigma-point schemes based on the third, fifth, seventh, and ninth order unscented transforms and the Gauss--Hermite quadrature rule. We compare the performance of the methods in two simulated experiments: a univariate nonlinear growth model as well as tracking of a maneuvering target. In the experiments, we also compare against approximate likelihood estimates obtained by particle filtering and extended Kalman filtering based methods. The experiments suggest that the higher-order unscented transforms may in some cases provide more accurate estimates
△ Less
Submitted 2 November, 2015; v1 submitted 23 April, 2015;
originally announced April 2015.
-
Hilbert Space Methods for Reduced-Rank Gaussian Process Regression
Authors:
Arno Solin,
Simo Särkkä
Abstract:
This paper proposes a novel scheme for reduced-rank Gaussian process regression. The method is based on an approximate series expansion of the covariance function in terms of an eigenfunction expansion of the Laplace operator in a compact subset of $\mathbb{R}^d$. On this approximate eigenbasis the eigenvalues of the covariance function can be expressed as simple functions of the spectral density…
▽ More
This paper proposes a novel scheme for reduced-rank Gaussian process regression. The method is based on an approximate series expansion of the covariance function in terms of an eigenfunction expansion of the Laplace operator in a compact subset of $\mathbb{R}^d$. On this approximate eigenbasis the eigenvalues of the covariance function can be expressed as simple functions of the spectral density of the Gaussian process, which allows the GP inference to be solved under a computational cost scaling as $\mathcal{O}(nm^2)$ (initial) and $\mathcal{O}(m^3)$ (hyperparameter learning) with $m$ basis functions and $n$ data points. Furthermore, the basis functions are independent of the parameters of the covariance function, which allows for very fast hyperparameter learning. The approach also allows for rigorous error analysis with Hilbert space theory, and we show that the approximation becomes exact when the size of the compact subset and the number of eigenfunctions go to infinity. We also show that the convergence rate of the truncation error is independent of the input dimensionality provided that the differentiability order of the covariance function is increases appropriately, and for the squared exponential covariance function it is always bounded by ${\sim}1/m$ regardless of the input dimensionality. The expansion generalizes to Hilbert spaces with an inner product which is defined as an integral over a specified input density. The method is compared to previously proposed methods theoretically and through empirical tests with simulated and real data.
△ Less
Submitted 1 August, 2019; v1 submitted 21 January, 2014;
originally announced January 2014.
-
Infinite-dimensional Bayesian filtering for detection of quasi-periodic phenomena in spatio-temporal data
Authors:
Arno Solin,
Simo Särkkä
Abstract:
This paper introduces a spatio-temporal resonator model and an inference method for detection and estimation of nearly periodic temporal phenomena in spatio-temporal data. The model is derived as a spatial extension of a stochastic harmonic resonator model, which can be formulated in terms of a stochastic differential equation (SDE). The spatial structure is included by introducing linear operator…
▽ More
This paper introduces a spatio-temporal resonator model and an inference method for detection and estimation of nearly periodic temporal phenomena in spatio-temporal data. The model is derived as a spatial extension of a stochastic harmonic resonator model, which can be formulated in terms of a stochastic differential equation (SDE). The spatial structure is included by introducing linear operators, which affect both the oscillations and dam**, and by choosing the appropriate spatial covariance structure of the driving time-white noise process. With the choice of the linear operators as partial differential operators, the resonator model becomes a stochastic partial differential equation (SPDE), which is compatible with infinite-dimensional Kalman filtering. The resulting infinite-dimensional Kalman filtering problem allows for a computationally efficient solution as the computational cost scales linearly with measurements in the temporal dimension. This framework is applied to weather prediction and to physiological noise elimination in fMRI brain data.
△ Less
Submitted 3 September, 2013; v1 submitted 11 March, 2013;
originally announced March 2013.
-
Room temperature ballistic transport in InSb quantum well nanodevices
Authors:
A. M. Gilbertson,
A. Kormanyos,
P. D. Buckle,
M. Fearn,
T. Ashley,
C. J. Lambert,
S. A. Solin,
L. F. Cohen
Abstract:
We report the room temperature observation of significant ballistic electron transport in shallow etched four-terminal mesoscopic devices fabricated on an InSb/AlInSb quantum well (QW) heterostructure with a crucial partitioned growth-buffer scheme. Ballistic electron transport is evidenced by a negative bend resistance signature which is quite clearly observed at 295 K and at current densities in…
▽ More
We report the room temperature observation of significant ballistic electron transport in shallow etched four-terminal mesoscopic devices fabricated on an InSb/AlInSb quantum well (QW) heterostructure with a crucial partitioned growth-buffer scheme. Ballistic electron transport is evidenced by a negative bend resistance signature which is quite clearly observed at 295 K and at current densities in excess of 10$^{6}$ A/cm$^{2}$. This demonstrates unequivocally that by using effective growth and processing strategies, room temperature ballistic effects can be exploited in InSb/AlInSb QWs at practical device dimensions.
△ Less
Submitted 21 November, 2011;
originally announced November 2011.
-
Ballistic transport and boundary scattering in InSb/InxAl1-xSb mesoscopic devices
Authors:
A. M. Gilbertson,
M. Fearn,
A. Kormányos,
D. E. Read,
C. J. Lambert,
M. T. Emeny,
T. Ashley,
S. A. Solin,
L. F. Cohen
Abstract:
We describe the influence of hard wall confinement and lateral dimension on the low temperature transport properties of long diffusive channels and ballistic crosses fabricated in an InSb/InxAl1-xSb heterostructure. Partially diffuse boundary scattering is found to play a crucial role in the electron dynamics of ballistic crosses and substantially enhance the negative bend resistance. Experimental…
▽ More
We describe the influence of hard wall confinement and lateral dimension on the low temperature transport properties of long diffusive channels and ballistic crosses fabricated in an InSb/InxAl1-xSb heterostructure. Partially diffuse boundary scattering is found to play a crucial role in the electron dynamics of ballistic crosses and substantially enhance the negative bend resistance. Experimental observations are supported by simulations using a classical billiard ball model for which good agreement is found when diffuse boundary scattering is included.
△ Less
Submitted 20 September, 2010;
originally announced September 2010.
-
Spin glassiness and power law scaling in a quasi-triangular spin-1/2 compound
Authors:
Jian Wu,
Julia S. Wildeboer,
Fletcher Werner,
Alexander Seidel,
Z. Nussinov,
S. A. Solin
Abstract:
We present data on the magnetic properties of two classes of layered spin S=1/2 antiferromagnetic quasi-triangular lattice materials: $Cu_{2(1-x)}Zn_{2x}(OH)_3NO_3$ ($0 < x < 0.65$) and its long chain organic derivatives $Cu_{2(1-x)}Zn_{2x}(OH)_3(C_7H_{15}COO)\cdot mH_2O$ ($0 < x < 0.29$), where non-magnetic Zn substitutes Cu isostructurally. It is found that the long-chain compounds, even in a cl…
▽ More
We present data on the magnetic properties of two classes of layered spin S=1/2 antiferromagnetic quasi-triangular lattice materials: $Cu_{2(1-x)}Zn_{2x}(OH)_3NO_3$ ($0 < x < 0.65$) and its long chain organic derivatives $Cu_{2(1-x)}Zn_{2x}(OH)_3(C_7H_{15}COO)\cdot mH_2O$ ($0 < x < 0.29$), where non-magnetic Zn substitutes Cu isostructurally. It is found that the long-chain compounds, even in a clean system in the absence of dilution, $x\!=\!0$, show spin-glass behavior, as evidenced by DC and AC susceptibility, and by time dependent magnetization measurements. A striking feature is the observation of a sharp crossover between two successive power law regimes in the DC susceptibility above the freezing temperature. Specific heat data are consistent with a conventional phase transition in the unintercalated compounds, and glassy behavior in the long chain compunds.
△ Less
Submitted 2 July, 2010;
originally announced July 2010.
-
Construction and Commissioning of the CALICE Analog Hadron Calorimeter Prototype
Authors:
C. Adloff,
Y. Karyotakis,
J. Repond,
A. Brandt,
H. Brown,
K. De,
C. Medina,
J. Smith,
J. Li,
M. Sosebee,
A. White,
J. Yu,
T. Buanes,
G. Eigen,
Y. Mikami,
O. Miller,
N. K. Watson,
J. A. Wilson,
T. Goto,
G. Mavromanolakis,
M. A. Thomson,
D. R. Ward,
W. Yan,
D. Benchekroun,
A. Hoummada
, et al. (205 additional authors not shown)
Abstract:
An analog hadron calorimeter (AHCAL) prototype of 5.3 nuclear interaction lengths thickness has been constructed by members of the CALICE Collaboration. The AHCAL prototype consists of a 38-layer sandwich structure of steel plates and highly-segmented scintillator tiles that are read out by wavelength-shifting fibers coupled to SiPMs. The signal is amplified and shaped with a custom-designed ASIC.…
▽ More
An analog hadron calorimeter (AHCAL) prototype of 5.3 nuclear interaction lengths thickness has been constructed by members of the CALICE Collaboration. The AHCAL prototype consists of a 38-layer sandwich structure of steel plates and highly-segmented scintillator tiles that are read out by wavelength-shifting fibers coupled to SiPMs. The signal is amplified and shaped with a custom-designed ASIC. A calibration/monitoring system based on LED light was developed to monitor the SiPM gain and to measure the full SiPM response curve in order to correct for non-linearity. Ultimately, the physics goals are the study of hadron shower shapes and testing the concept of particle flow. The technical goal consists of measuring the performance and reliability of 7608 SiPMs. The AHCAL was commissioned in test beams at DESY and CERN. The entire prototype was completed in 2007 and recorded hadron showers, electron showers and muons at different energies and incident angles in test beams at CERN and Fermilab.
△ Less
Submitted 12 March, 2010;
originally announced March 2010.
-
Finite element modeling of extraordinary optoconductance in GaAs-In metal-semiconductor hybrid structures
Authors:
K. A. Wieland,
Yun Wang,
S. A. Solin,
A. M. Girgis,
L. R. Ram-Mohan
Abstract:
We present a detailed discussion of extraordinary optoconductance (EOC). Experimental data was acquired via macroscopic metal-semiconductor hybrid structures composed of GaAs and In and subjected to illumination from an Ar ion laser. A drift diffusion model using the finite element method (FEM) provided a reasonable fit to the data. EOC is explored as a function of laser position, bias current,…
▽ More
We present a detailed discussion of extraordinary optoconductance (EOC). Experimental data was acquired via macroscopic metal-semiconductor hybrid structures composed of GaAs and In and subjected to illumination from an Ar ion laser. A drift diffusion model using the finite element method (FEM) provided a reasonable fit to the data. EOC is explored as a function of laser position, bias current, laser power density, and temperature. The positional dependence of the voltage is accounted for by the Dember effect, with the model incorporating the excess hole distribution based on the carrier mobility, and thus the mean free path. The bias current is found to produce a linear voltage offset and does not influence the EOC. A linear relationship is found between the laser power density and the voltage in the bare and hybrid devices. This dependence is reproduced in the model by a generation rate parameter which is related to the power density. Incorporating the mobility and diffusion temperature dependence, the model directly parallels the temperature dependence of the EOC without the use of fitting parameters.
△ Less
Submitted 9 March, 2006; v1 submitted 6 February, 2006;
originally announced February 2006.
-
Design and Properties of a scanning EMR probe Microscope
Authors:
S. A. Solin
Abstract:
The design, fabrication, and predicted performance of a new type of magnetic scanning probe microscope based on the newly discovered phenomenon of extraordinary magnetoresistance (EMR) is described. It is shown that the new probe should advance the state of the art of both sensitivity and spatial resolution by an order of magnitude or more.
The design, fabrication, and predicted performance of a new type of magnetic scanning probe microscope based on the newly discovered phenomenon of extraordinary magnetoresistance (EMR) is described. It is shown that the new probe should advance the state of the art of both sensitivity and spatial resolution by an order of magnitude or more.
△ Less
Submitted 6 February, 2006;
originally announced February 2006.
-
Carrier Mobilities in Delta-doped Heterostructures
Authors:
Y. Shao,
S. A. Solin,
L. R. Ram-Mohan
Abstract:
For applications to sensor design, the product nxmu of the electron density n and the mobility mu is a key parameter to be optimized for enhanced device sensitivity. We model the carrier mobility in a two dimensional electron gas (2DEG) layer developed in a delta-doped heterostructure. The subband energy levels, electron wave functions, and the band-edge profile are obtained by numerically solvi…
▽ More
For applications to sensor design, the product nxmu of the electron density n and the mobility mu is a key parameter to be optimized for enhanced device sensitivity. We model the carrier mobility in a two dimensional electron gas (2DEG) layer developed in a delta-doped heterostructure. The subband energy levels, electron wave functions, and the band-edge profile are obtained by numerically solving the Schrodinger and Poisson equations self-consistently. The electron mobility is calculated by including contributions of scattering from ionized impurities, the background neutral impurities, the deformation potential acoustic phonons, and the polar optical phonons. We calculate the dependencies of nxmu on temperature, spacer layer thickness, do** density, and the quantum well thickness. The model is applied to delta-doped quantum well heterostructures of AlInSb-InSb. At low temperature, mobilities as high as 1.3x10^3 m^2/Vs are calculated for large spacer layers (400 A) and well widths (400 A). The corresponding room temperature mobility is 10 m^2/Vs. The dependence of nxmu shows a maximum for a spacer thickness of 300 A for higher background impurity densities while it continues to increase monotonically for lower background impurity densities; this has implications for sensor design.
△ Less
Submitted 6 February, 2006;
originally announced February 2006.
-
The Effect of Transfer Printing on Pentacene Thin-Film Crystal Structure
Authors:
Y. Shao,
S. A. Solin,
D. R. Hines,
E. D. Williams
Abstract:
The thermal deposition and transfer Printing method had been used to produce pentacene thin-films on SiO2/Si and plastic substrates (PMMA and PVP), respectively. X-ray diffraction patterns of pentacene thin films showed reflections associated with highly ordered polycrystalline films and a coexistence of two polymorph phases classified by their d-spacing, d(001): 14.4 and 15.4 A.The dependence o…
▽ More
The thermal deposition and transfer Printing method had been used to produce pentacene thin-films on SiO2/Si and plastic substrates (PMMA and PVP), respectively. X-ray diffraction patterns of pentacene thin films showed reflections associated with highly ordered polycrystalline films and a coexistence of two polymorph phases classified by their d-spacing, d(001): 14.4 and 15.4 A.The dependence of the c-axis correlation length and the phase fraction on the film thickness and printing temperature were measured. A transition from the 15.4 A phase towards 14.4 A phase was also observed with increasing film thickness. An increase in the c-axis correlation length of approximately 12% ~16% was observed for Pn films transfer printed onto a PMMA coated PET substrate at 100~120 C as compared to as-grown Pn films on SiO2/Si substrates. The transfer printing method is shown to be an attractive for the fabrication of pentacene thin-film transistors on flexible substrates partly because of the resulting improvement in the quality of the pentacene film.
△ Less
Submitted 6 February, 2006;
originally announced February 2006.
-
Extraordinary Phenomena in Semiconductor-Metal Hybrid Nanostructures Based on Bilinear Conformal Map**
Authors:
S. A. Solin
Abstract:
We have shown that bilinear conformal map** can be used to transform 4-lead internally shunted EMR semiconductor-metal hybrid structures to galvanomagnetically equivalent externally shunted 4 lead structures. The latter are compatible with the fabrication of nanoscale EMR devices while the former are not. Mapped rectangular EMR van der Pauw plate exhibit very large EMR values in both macroscop…
▽ More
We have shown that bilinear conformal map** can be used to transform 4-lead internally shunted EMR semiconductor-metal hybrid structures to galvanomagnetically equivalent externally shunted 4 lead structures. The latter are compatible with the fabrication of nanoscale EMR devices while the former are not. Mapped rectangular EMR van der Pauw plate exhibit very large EMR values in both macroscopic and nanoscopic form. We have also shown that the map** procedure applied in the case of EMR will also be applicable to other generalized EXX structures.
△ Less
Submitted 6 February, 2006;
originally announced February 2006.
-
Position-sensitive detector for the 6-meter optical telescope
Authors:
V. Debur,
T. Arkhipova,
G. Beskin,
V. Plokhotnichenko,
M. Pakhomov,
M. Smirnova,
A. Solin
Abstract:
The Position-Sensitive Detector (PSD) for photometrical and spectral observation on the 6-meter optical telescope of the Special Astrophysical Observatory (Russia) is described. The PSD consists of a position-sensitive tube, amplifiers of output signals, analog-to-digital converters (ADC) and a digital logic plate, which produces a signal for ADC start and an external strob pulse for reading inf…
▽ More
The Position-Sensitive Detector (PSD) for photometrical and spectral observation on the 6-meter optical telescope of the Special Astrophysical Observatory (Russia) is described. The PSD consists of a position-sensitive tube, amplifiers of output signals, analog-to-digital converters (ADC) and a digital logic plate, which produces a signal for ADC start and an external strob pulse for reading information by registration system. If necessary, the thermoelectric cooler can be used. The position-sensitive tube has the following main elements: a photocathode, electrodes of inverting optics, a block of microchannel plates (MCP) and a position-sensitive collector of quadrant type. The main parameters of the PSD are the diameter of the sensitive surface is 25 mm, the spatial resolution is better than 100 (μ)m in the centre and a little worse on the periphery; the dead time is near 0.5 (μ)s; the detection quantum efficiency is defined by the photocathode and it is not less than 0.1, as a rule; dark current is about hundreds of cps, or less, when cooling. PSD spectral sensitivity depends on the type of photocathode and input window material. We use a multialkali photocathode and a fiber or UV-glass, which gives the short- wave cut of 360 nm or 250 nm, respectively.
△ Less
Submitted 15 October, 2003; v1 submitted 14 October, 2003;
originally announced October 2003.
