Using the Projected Belief Network at High Dimensions

IEEE transactions on neural networks and learning systems, 2019

This paper addresses the duality between deterministic feed-forward neural networks (FF-NNs), and linear Bayesian networks (LBNs), which are generative stochastic models representing probability distributions over the visible data based on a linear function of a set of latent (hidden) variables. The maximum entropy principle is used to define a unique generative model corresponding to each FF-NN, called projected belief network (PBN). The FF-NN exactly recovers the hidden variables of the dual PBN. The large-◆ asymptotic approximation to the PBN has the familiar structure of an LBN, with the addition of an invertible non-linear transformation operating on the latent variables. It is shown that the exact nature of the PBN depends on the range of the input (visible) data-details for three cases of input data range are provided. The likelihood function of the PBN is straightforward to calculate, allowing it to be used as a generative classifier. An example is provided in which a generative classifier based on the PBN has comparable performance to a deep belief network in classifying handwritten characters. In addition, several examples are provided that demonstrate the duality relationship, for example by training networks from either side of the duality.

IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2017

Automatic sound event recognition (SER) has recently attracted renewed interest. Although practical SER system has many useful applications in everyday life, SER is challenging owing to the variations among sounds and noises in the realworld environment. This work presents a novel feature extraction and classification method to solve the problem of sound event recognition. An audiovisual descriptor, called the auditoryreceptive-field binary pattern (ARFBP), is designed based on the spectrogram image feature (SIF), the cepstral features, and the human auditory receptive field model. The extracted features are then fed into a classifier to perform event classification. The proposed classifier, called the hierarchical-diving deep belief network (HDDBN), is a deep neural network (DNN) system that hierarchically learns the discriminative characteristics from physical feature representation to the abstract concept. The performance of our proposed system was verified using several experiments under various conditions. Using the RWCP dataset, the proposed system achieved a recognition rate of 99.27% for real-world sound data in 105 categories. Under noisy conditions, the developed system is very robust, with which it achieved 95.06% recognition rate with 0 dB signal-to-noise ratio (SNR). Using the TUT sound event dataset, the proposed system achieves error rates of 0.81 and 0.73 in sound event detection in home and residential area scenes. The experimental results reveal that the proposed system outperformed the other systems in this field. Index Terms-Environmental sound, spectrogram image feature, auditory receptive fields binary patterns, hierarchical diving deep belief network.

This paper evaluates neural network (NN) based systems and compares them to Gaussian mixture model (GMM) and hidden Markov model (HMM) approaches for acoustic scene classification (SC) and polyphonic acoustic event detection (AED) that are applied to data of the "Detection and Classification of Acoustic Scenes and Events 2016" (DCASE'16) challenge, task 1 and task 3, respectively. For both tasks, the use of deep neural networks (DNNs) and features based on an amplitude modulation filterbank and a Gabor filterbank (GFB) are evaluated and compared to standard approaches. For SC, additionally a time-delay NN approach is proposed that enables analysis of long contextual information similar to recurrent NNs but with training efforts comparable to conventional DNNs. The SC system proposed for task 1 of the DCASE'16 challenge attains a recognition accuracy of 77.5%, which is 5.6% higher compared to the DCASE'16 baseline system. For the AED task, DNNs are adopted in tandem and hybrid approaches, i.e., as part of HMM-based systems. These systems are evaluated for the polyphonic data of task 3 from the DCASE'16 challenge. Several strategies to address the issue of polyphony are considered. It is shown that DNN-based systems perform less accurate than the traditional systems for this task. Best results are achieved using GFB features in combination with a multiclass GMM-HMM back end.

In music, polyphonic describes as a musical texture when two or more notes are playing at the same time. Polyphonic Music Transcription is the processes of automatically transcribing notes from an input audio signal. This study discusses state of art methods in polyphonic music transcription. In general ,music transcription techniques can be categorised into 3 group Signal Processing Methods Statistical Modelling base Method and Machine Learning Methods. This study focuses on machine learning method. Recently unsupervised feature learning methods such as RBM(Restricted Boltzman machine ) have shown great promise as a way of extracting features from high dimensional data, such as image or audio. In this paper, we applied deep belief networks to musical data and evaluate the learned feature representations on polyphonic piano transcription. A polyphonic music transcription system was developed and evaluated on a public piano data set named MAPS using metric of accuracy. The result on a small data set shows 44.2% accuracy which is promising for further investigation of this method on a larger data set.

International Journal for Research in Applied Science and Engineering Technology IJRASET, 2020

It is a difficult task of continuous automatic speech recognition, translating of spoken words into text due to the excessive viability in speech signals. In recent years speech recognition has been accomplishing pinnacle of success however it still has few limitations to overcome. Deep learning also known as representation learning or sometimes referred as unsupervised feature learning, is a subset of machine learning. Deep learning is becoming a conventional technology for speech recognition and has efficiently replaced Gaussian mixtures for speech recognition on a global scale. The predominant goal of this undertaking is to apply deep learning algorithms, together with Deep Neural Networks (DNN) and Deep Belief Networks (DBN), for automatic non-stop speech recognition. Keywords: Gaussian Mixture Model (GMM), Hidden Markov Models (HMMs), Deep Neural Networks (DNN), Deep Belief Networks (DBN).

Journal of the Acoustical Society of America, 2021

Deep learning models have become potential candidates for auditory neuroscience research, thanks to their recent successes in a variety of auditory tasks, yet these models often lack interpretability to fully understand the exact computations that have been performed. Here, we proposed a parametrized neural network layer, which computes specific spectro-temporal modulations based on Gabor filters [learnable spectro-temporal filters (STRFs)] and is fully interpretable. We evaluated this layer on speech activity detection, speaker verification, urban sound classification, and zebra finch call type classification. We found that models based on learnable STRFs are on par for all tasks with state-of-the-art and obtain the best performance for speech activity detection. As this layer remains a Gabor filter, it is fully interpretable. Thus, we used quantitative measures to describe distribution of the learned spectro-temporal modulations. Filters adapted to each task and focused mostly on low temporal and spectral modulations. The analyses show that the filters learned on human speech have similar spectro-temporal parameters as the ones measured directly in the human auditory cortex. Finally, we observed that the tasks organized in a meaningful way: the human vocalization tasks closer to each other and bird vocalizations far away from human vocalizations and urban sounds tasks. V

We show how to use "complementary priors" to eliminate the explaining away effects that make inference difficult in densely-connected belief nets that have many hidden layers. Using complementary priors, we derive a fast, greedy algorithm that can learn deep, directed belief networks one layer at a time, provided the top two layers form an undirected associative memory. The fast, greedy algorithm is used to initialize a slower learning procedure that fine-tunes the weights using a contrastive version of the wake-sleep algorithm. After fine-tuning, a network with three hidden layers forms a very good generative model of the joint distribution of handwritten digit images and their labels. This generative model gives better digit classification than the best discriminative learning algorithms. The low-dimensional manifolds on which the digits lie are modelled by long ravines in the free-energy landscape of the top-level associative memory and it is easy to explore these ravines by using the directed connections to display what the associative memory has in mind.

2016

Learning representation from audio data has shown advantages over the hand-crafted features such as Mel Frequency Cepstral Coefficients (MFCC) in many audio applications. In most of the representation learning approaches, the connectionist systems have been used to learn and extract latent features from the fixed length data. In this paper, we propose an approach to combine the learned features and the MFCC features for speaker recognition task, which can be applied to audio scripts of different length. In particular, we study the use of features from different levels of Deep Belief Network for quantizing the audio data into vectors of audio-word counts. These vectors represent the audio scripts of different length that make them easier to train a classifier. We show in the experiment that the audio-word count vectors generated from mixture of DBN features at different layers give better performance than the MFCC features. We also can achieve further improvement by combining the audio-word count vector and the MFCC features.

Neural Computation, 2008

Deep Belief Networks (DBN) are generative neural network models with many layers of hidden explanatory factors, recently introduced by Hinton et al., along with a greedy layer-wise unsupervised learning algorithm. The building block of a DBN is a probabilistic model called a Restricted Boltzmann Machine (RBM), used to represent one layer of the model. Restricted Boltzmann Machines are interesting because inference is easy in them, and because they have been successfully used as building blocks for training deeper models. We first prove that adding hidden units yields strictly improved modeling power, while a second theorem shows that RBMs are universal approximators of discrete distributions. We then study the question of whether DBNs with more layers are strictly more powerful in terms of representational power. This suggests a new and less greedy criterion for training RBMs within DBNs.

We provide the first direct comparison of sum-product networks (SPNs) and deep-belief networks on speech, and the first application of SPNs to acoustic-articulatory inversion. Interestingly, speech from individuals with cerebral palsy is reconstructed significantly more accurately across all manners of articulation using SPNs than when using DBNs. In order to select appropriate input parameters, we first compare MFCCs, wavelets, scattering coefficients, and vocal 'tract variables' as predictor variables to phonological features. Here, MFCCs provide for more accurate classification over a broad array of phonological categories (in the high 90s in many cases) than the other feature types. All experiments use the MOCHA-TIMIT and TORGO acoustic-articulatory databases.