Peak detection and localization in a noisy signal with an unknown baseline is a fundamental task in signal processing applications such as spectroscopy. A current trend in signal processing is to reformulate traditional processing pipelines as (deep) neural networks that can be trained end-to-end. A trainable algorithm for baseline removal and peak localization can serve as an important module in such a processing pipeline. In practical applications, one of the most successful approaches to joint baseline suppression and peak localization is based on the continuous wavelet transform: We re-formulate this as a convolutional neural network (CNN) followed by a non-linear readout layer. On a synthetic benchmark we demonstrate that with sufficient training data, the CNN approach consistently outperforms the optimized continuous wavelet method by means of adapting to the spectral peak shape, noise level, and characteristics of the baseline. The CNN approach to peak localization shows great promise, as it can more efficiently leverage data to outperform the current state of the art, and can readily be extended and incorporated as a module in a larger neural network architecture.

Raman spectroscopy is a well-known analytical technique for identifying and analyzing chemical species. Since Raman scattering is a weak effect, surface-enhanced Raman spectroscopy (SERS) is often employed to amplify the signal. SERS signal surface mapping is a common method for detecting trace amounts of target molecules. Since the method produce large amounts of data and, in the case of very low concentrations, low signal-to-noise (SNR) ratio, ability to extract relevant spectral features is crucial. We propose a pseudo-Voigt model as a constrained source separation model, that is able to directly and reliably identify the Raman modes, with overall performance similar to the state of the art non-negative matrix factorization approach. However, the model provides better interpretation and is a step towards enabling the use of SERS in detection of trace amounts of molecules in real-life settings.

Personalization of multi-parameter hearing aids involves an initial fitting followed by a manual knowledge-based trial-and-error fine-tuning from ambiguous verbal user feedback. The result is an often suboptimal HA setting whereby the full potential of modern hearing aids is not utilized. This article proposes an interactive hearing-aid personalization system that obtains an optimal individual setting of the hearing aids from direct perceptual user feedback. Results obtained with ten hearing-impaired subjects show that ten to twenty pairwise user assessments between different settings-equivalent to 5-10 min-is sufficient for personalization of up to four hearing-aid parameters. A setting obtained by the system was significantly preferred by the subject over the initial fitting, and the obtained setting could be reproduced with reasonable precision. The system may have potential for clinical usage to assist both the hearing-care professional and the user.

An adversary is an agent designed to make a classification system perform in some particular way, e.g., increase the probability of a false negative. Recent work builds adversaries for deep learning systems applied to image object recognition, exploiting the parameters of the system to find the minimal perturbation of the input image such that the system misclassifies it with high confidence. We adapt this approach to construct and deploy an adversary of deep learning systems applied to music content analysis. In our case, however, the system inputs are magnitude spectral frames, which require special care in order to produce valid input audio signals from network- derived perturbations . For two different train-test partitionings of two benchmark datasets, and two different architectures , we find that this adversary is very effective. We find that convolutional architectures are more robust compared to systems based on a majority vote over individually classified audio frames. Furthermore , we experiment with a new system that integrates an adversary into the training loop, but do not find that this improves the resilience of the system to new adversaries.

A fundamental and general representation of audio and music which integrates multi-modal data sources is important for both application and basic research purposes. In this paper we address this challenge by proposing a multi-modal version of the Latent Dirichlet Allocation model which provides a joint latent representation. We evaluate this representation on the Million Song Dataset by integrating three fundamentally different modalities, namely tags, lyrics, and audio features. We show how the resulting representation is aligned with common 'cognitive' variables such as tags, and provide some evidence for the common assumption that genres form an acceptable categorization when evaluating latent representations of music. We furthermore quantify the model by its predictive performance in terms of genre and style, providing benchmark results for the Million Song Dataset.

We present a colorimetric sensor array which is able to detect explosives such as DNT, TNT, HMX, RDX and TATP and identifying volatile organic compounds in the presence of water vapor in air. To analyze colorimetric sensors with statistical methods, a suitable representation of sensory readings is required. We present a new approach of extracting features from a colorimetric sensor array based on a color distribution representation. For each sensor in the array, we construct a K-nearest neighbor classifier based on the Hellinger distances between color distribution of a test compound and the color distribution of all the training compounds. The performance of this set of classifiers are benchmarked against a set of K-nearest neighbor classifiers that is based on traditional feature representation (e.g., mean or global mode). The suggested approach of using the entire distribution outperforms the traditional approaches which use a single feature.

Music recommendation is an important aspect of many streaming services and multi-media systems, however, it is typically based on so-called collaborative filtering methods. In this paper we consider the recommendation task from a personal viewpoint and examine to which degree music preference can be elicited and predicted using simple and robust queries such as pairwise comparisons. We propose to model - and in turn predict - the pairwise music preference using a very flexible model based on Gaussian Process priors for which we describe the required inference. We further propose a specific covariance function and evaluate the predictive performance on a novel dataset. In a recommendation style setting we obtain a leave-one-out accuracy of 74% compared to 50% with random predictions, showing potential for further refinement and evaluation.

In this paper, we present a new evaluation approach for missing data techniques (MDTs) where the efficiency of those are investigated using listwise deletion method as reference. We experiment on classification problems and calculate misclassification rates (MR) for different missing data percentages (MDP) using a missing completely at random (MCAR) scheme. We compare three MDTs: pairwise deletion (PW), mean imputation (MI) and a maximum likelihood method that we call complete expectation maximization (CEM). We use a synthetic dataset, the Iris dataset and the Pima Indians Diabetes dataset. We train a Gaussian mixture model (GMM). We test the trained GMM for two cases, in which test dataset is missing or complete. The results show that CEM is the most efficient method in both cases while MI is the worst performer of the three. PW and CEM proves to be more stable, in particular for higher MDP values than MI.

In this paper we describe experiments carried out at the Nibe music festival in Denmark involving the use of mobile phones to augment the participants' sound experience at the concerts. The experiments involved N=19 test participants that used a mobile phone with a headset playing back sound received over FM from the PA audio mixer system. Based on the location of the participant (distance to the stage) a delay was estimated and introduced to the playback on the mobile phone in order to align the sound in the headset with that from the on-stage speakers. We report our findings from our initial "in-the-wild" experiments augmenting the sound experience at two concerts at this music festival.

Separation of speech mixtures, often referred to as the cocktail party problem, has been studied for decades. In many source separation tasks, the separation method is limited by the assumption of at least as many sensors as sources. Further, many methods require that the number of signals within the recorded mixtures be known in advance. In many real-world applications, these limitations are too restrictive. We propose a novel method for underdetermined blind source separation using an instantaneous mixing model which assumes closely spaced microphones. Two source separation techniques have been combined, independent component analysis (ICA) and binary time - frequency (T-F) masking. By estimating binary masks from the outputs of an ICA algorithm, it is possible in an iterative way to extract basis speech signals from a convolutive mixture. The basis signals are afterwards improved by grouping similar signals. Using two microphones, we can separate, in principle, an arbitrary number of mixed speech signals. We show separation results for mixtures with as many as seven speech signals under instantaneous conditions. We also show that the proposed method is applicable to segregate speech signals under reverberant conditions, and we compare our proposed method to another state-of-the-art algorithm. The number of source signals is not assumed to be known in advance and it is possible to maintain the extracted signals as stereo signals.

Temporal feature integration is the process of combining all the feature vectors in a time window into a single feature vector in order to capture the relevant temporal information in the window. The mean and variance along the temporal dimension are often used for temporal feature integration, but they capture neither the temporal dynamics nor dependencies among the individual feature dimensions. Here, a multivariate autoregressive feature model is proposed to solve this problem for music genre classification. This model gives two different feature sets, the diagonal autoregressive (DAR) and multivariate autoregressive (MAR) features which are compared against the baseline mean-variance as well as two other temporal feature integration techniques. Reproducibility in performance ranking of temporal feature integration methods were demonstrated using two data sets with five and eleven music genres, and by using four different classification schemes. The methods were further compared to human performance. The proposed MAR features perform better than the other features at the cost of increased computational complexity.

A general variational Bayesian framework for iterative data and parameter estimation for coherent detection is introduced as a generalization of the EM-algorithm. Explicit solutions are given for MIMO channel estimation with Gaussian prior and noise covariance estimation with inverse-Wishart prior. Simulation of a GSM-like system provides empirical proof that the VBEM-algorithm is able to provide better performance than the EM-algorithm. However, if the posterior distribution is highly peaked, the VBEM-algorithm approaches the EM-algorithm and the gain disappears. The potential gain is therefore greatest in systems with a small amount of observations compared to the number of parameters to be estimated

Spectral clustering methods were proposed recently which rely on the eigenvalue decomposition of an affinity matrix. In this letter, the affinity matrix is created from the elements of a nonparametric density estimator and then decomposed to obtain posterior probabilities of class membership. Hyperparameters are selected using standard cross-validation methods.

Many different short-time features, using time windows of 10-30 ms, have been proposed for music segmentation, retrieval and genre classification. However, often the available time frame of the music to make the actual decision or comparison (the decision time horizon) is in the range of seconds instead of milliseconds. The problem of making new features on the larger time scale from the short-time features (feature integration) has received only little attention. The paper investigates different methods for feature integration and late information fusion for music genre classification. A new feature integration technique, the AR model, is proposed and seemingly outperforms the commonly used mean-variance features.

When the ICA source separation problem is solved by maximum likelihood, a proper choice of the parameters is important. A comparison has been performed between the use of a mixing matrix and the use of the separation matrix as parameters in the likelihood. By looking at a general behavior of the cost function as function of the mixing matrix or as function of the separation matrix, it is explained and illustrated why it is better to select the separation matrix as a parameter than to use the mixing matrix as a parameter. The behavior of the natural gradient in the two cases has been considered as well as the influence of pre-whitening.

Language independent 'bag-of-words' representations are surprisingly effective for text classification. The representation is high dimensional though, containing many non-consistent words for text categorization. These non-consistent words result in reduced generalization performance of subsequent classifiers, e.g., from ill-posed principal component transformations. In this communication, our aim is to study the effect of reducing the least relevant words from the bag-of-words representation. We consider a new approach, using neural network based sensitivity maps and information gain for determination of term relevancy, when pruning the vocabularies. With reduced vocabularies, documents are classified using a latent semantic indexing representation and a probabilistic neural network classifier. Reducing the bag-of-words vocabularies with 90%-98%, we find consistent classification improvement using two mid size data-sets. We also study the applicability of information gain and sensitivity maps for automated keyword generation.

The object of Bayesian modelling is predictive distribution, which, in a forecasting scenario, enables evaluation of forecasted values and their uncertainties. We focus on reliably estimating the predictive mean and variance of forecasted values using Bayesian kernel based models such as the Gaussian process and the relevance vector machine. We derive novel analytic expressions for the predictive mean and variance for Gaussian kernel shapes under the assumption of a Gaussian input distribution in the static case, and of a recursive Gaussian predictive density in iterative forecasting. The capability of the method is demonstrated for forecasting of time-series and compared to approximate methods.

We extend MacKays Bayesian approach to neural classifiers to include an outlier detector mechanism. We show that the outlier detector can locate misclassified samples.

In certain applications of independent component analysis (ICA) it is of interest to test hypotheses concerning the number of components or simply to test whether a given number of components is significant relative to a "white noise" null hypothesis. We estimate probabilities of such competing hypotheses for ICA based on dynamic decorrelation. The probabilities are evaluated in the so-called Bayesian information criterion approximation, however, they are able to detect the content of dynamic components as efficiently as an unbiased test set estimator.

We apply and discuss generalizable Gaussian mixture (GGM) models for text mining. The model automatically adapts model complexity for a given text representation. We show that the generalizability of these models depends on the dimensionality of the representation and the sample size. We discuss the relation between supervised and unsupervised learning in the test data. Finally, we implement a novelty detector based on the density model.

This paper describes an underwater object discrimination system applied to recognize cylinders of various compositions from different angles. The system is based on a new combination of simulated dolphin clicks, simulated auditory filters and artificial neural networks. The model demonstrates its potential on real data collected from four different cylinders in an environment where the angles were controlled in order to evaluate the models capabilities to recognize cylinders independent of angles.

Andersen et al. (1997) and Larsen et al. (1996, 1997) suggested a regularization scheme which iteratively adapts regularization parameters by minimizing validation error using simple gradient descent. In this contribution we present an improved algorithm based on the conjugate gradient technique. Numerical experiments with feedforward neural networks successfully demonstrate improved generalization ability and lower computational cost.

This paper addresses a new framework for designing robust neural network classifiers. The network is optimized using the maximum a posteriori technique, i.e., the cost function is the sum of the log-likelihood and a regularization term (prior). In order to perform robust classification, we present a modified likelihood function which incorporates the potential risk of outliers in the data. This leads to the introduction of a new parameter, the outlier probability. Designing the neural classifier involves optimization of network weights as well as outlier probability and regularization parameters. We suggest to adapt the outlier probability and regularisation parameters by minimizing the error on a validation set, and a simple gradient descent scheme is derived. In addition, the framework allows for constructing a simple outlier detector. Experiments with artificial data demonstrate the potential of the suggested framework.

This paper addresses the problem of generalization error estimation in neural networks. A new early stop criterion based on a Bootstrap estimate of the generalization error is suggested. The estimate does not require the network to be trained to the minimum of the cost function, as required by other methods based on asymptotic theory. Moreover, in contrast to methods based on cross-validation which require data left out for testing, and thus biasing the estimate, the Bootstrap technique does not have this disadvantage. The potential of the suggested technique is demonstrated on various time-series problems.

This article presents a generalization of the Boltzmann machine that allows us to use the learning rule for a much wider class of maximum likelihood and maximum a posteriori problems, including both supervised and unsupervised learning. Furthermore, the approach allows us to discuss regularization and generalization in the context of Boltzmann machines. We provide an illustrative example concerning parameter estimation in an inhomogeneous Markov field. The regularized adaptation produces a parameter set that closely resembles the "teacher" parameters, hence, will produce segmentations that closely reproduce those of the inhomogeneous teacher network.

A new algorithm for successive identification of seismic reflections is proposed. Generally, the algorithm can be viewed as a curve matching method for images with specific structure. However, in the paper, the algorithm works on seismic signals assembled to constitute an image in which the investigated reflections produce curves. In numerical examples, the authors work on signals assembled in CMP gathers. The key idea of the algorithm is to estimate the reflection curve parameters and the reflection coefficients along these curves by combining the multipulse technique and the generalized Radon transform. The multipulse technique is used for wavelet identification in each trace, and the generalized Radon transform is used to coordinate the wavelet identification between the individual traces. Furthermore, a stop criterion and a reflection validation procedure are presented. The stop criterion stops the reflection estimation when the actual estimated reflection is insignificant. The reflection validation procedure ensures that the estimated reflections follow the shape of the investigated reflection curves. The algorithm is successfully used in two numerical examples. One is based on a synthetic CMP gather, whereas the other is based on a real recorded CMP gather. Initially, the algorithm requires an estimate of the wavelet that can be performed by any wavelet estimation method.