Spiking neural networks are the basis of versatile and power-efficient information processing in the brain. Although we currently lack a detailed understanding of how these networks compute, recently developed optimization techniques allow us to instantiate increasingly complex functional spiking neural networks in-silico. These methods hold the promise to build more efficient non-von-Neumann computing hardware and will offer new vistas in the quest of unraveling brain circuit function. To accelerate the development of such methods, objective ways to compare their performance are indispensable. Presently, however, there are no widely accepted means for comparing the computational performance of spiking neural networks. To address this issue, we introduce two spike-based classification datasets, broadly applicable to benchmark both software and neuromorphic hardware implementations of spiking neural networks. To accomplish this, we developed a general audio-to-spiking conversion procedure inspired by neurophysiology. Further, we applied this conversion to an existing and a novel speech dataset. The latter is the free, high-fidelity, and word-level aligned Heidelberg digit dataset that we created specifically for this study. By training a range of conventional and spiking classifiers, we show that leveraging spike timing information within these datasets is essential for good classification accuracy. These results serve as the first reference for future performance comparisons of spiking neural networks.
Saving rainforests is a key to halting adverse climate changes. In this paper, we introduce an innovative solution built on acoustic surveillance and machine learning technologies to help rainforest conservation. In particular, We propose new convolutional neural network (CNN) models for environmental sound classification and achieved promising preliminary results on two datasets, including a public audio dataset and our real rainforest sound dataset. The proposed audio classification models can be easily extended in an automated machine learning paradigm and integrated in cloud-based services for real world deployment.
We demonstrate the existence of universal adversarial perturbations, which can fool a family of audio classification architectures, for both targeted and untargeted attack scenarios. We propose two methods for finding such perturbations. The first method is based on an iterative, greedy approach that is well-known in computer vision: it aggregates small perturbations to the input so as to push it to the decision boundary. The second method, which is the main contribution of this work, is a novel penalty formulation, which finds targeted and untargeted universal adversarial perturbations. Differently from the greedy approach, the penalty method minimizes an appropriate objective function on a batch of samples. Therefore, it produces more successful attacks when the number of training samples is limited. Moreover, we provide a proof that the proposed penalty method theoretically converges to a solution that corresponds to universal adversarial perturbations. We also demonstrate that it is possible to provide successful attacks using the penalty method when only one sample from the target dataset is available for the attacker. Experimental results on attacking various 1D CNN architectures have shown attack success rates higher than 85.0% and 83.1% for targeted and untargeted attacks, respectively using the proposed penalty method.
Long-range dependencies modeling, widely used in capturing spatiotemporal correlation, has shown to be effective in CNN dominated computer vision tasks. Yet neither stacks of convolutional operations to enlarge receptive fields nor recent nonlocal modules is computationally efficient. In this paper, we present a generic family of lightweight global descriptors for modeling the interactions between positions across different dimensions (e.g., channels, frames). This descriptor enables subsequent convolutions to access the informative global features with negligible computational complexity and parameters. Benchmark experiments show that the proposed method can complete state-of-the-art long-range mechanisms with a significant reduction in extra computing cost. Code available at https://github.com/HolmesShuan/Compact-Global-Descriptor.
Detecting sleepiness from spoken language is an ambitious task, which is addressed by the Interspeech 2019 Computational Paralinguistics Challenge (ComParE). We propose an end-to-end deep learning approach to detect and classify patterns reflecting sleepiness in the human voice. Our approach is based solely on a moderately complex deep neural network architecture. It may be applied directly on the audio data without requiring any specific feature engineering, thus remaining transferable to other audio classification tasks. Nevertheless, our approach performs similar to state-of-the-art machine learning models.
Residual learning is a recently proposed learning framework to facilitate the training of very deep neural networks. Residual blocks or units are made of a set of stacked layers, where the inputs are added back to their outputs with the aim of creating identity mappings. In practice, such identity mappings are accomplished by means of the so-called skip or residual connections. However, multiple implementation alternatives arise with respect to where such skip connections are applied within the set of stacked layers that make up a residual block. While ResNet architectures for image classification using convolutional neural networks (CNNs) have been widely discussed in the literature, few works have adopted ResNet architectures so far for 1D audio classification tasks. Thus, the suitability of different residual block designs for raw audio classification is partly unknown. The purpose of this paper is to analyze and discuss the performance of several residual block implementations within a state-of-the-art CNN-based architecture for end-to-end audio classification using raw audio waveforms. For comparison purposes, we analyze as well the performance of the residual blocks under a similar 2D architecture using a conventional time-frequency audio represen-tation as input. The results show that the achieved accuracy is considerably dependent, not only on the specific residual block implementation, but also on the selected input normalization.
Stochastic variational inference for Bayesian deep neural network (DNN) requires specifying priors and approximate posterior distributions over neural network weights. Specifying meaningful weight priors is a challenging problem, particularly for scaling variational inference to deeper architectures involving high dimensional weight space. We propose MOdel Priors with Empirical Bayes using DNN (MOPED) method to choose informed weight priors in Bayesian neural networks. We formulate a two-stage hierarchical modeling, first find the maximum likelihood estimates of weights with DNN, and then set the weight priors using empirical Bayes approach to infer the posterior with variational inference. We empirically evaluate the proposed approach on real-world tasks including image classification, video activity recognition and audio classification with varying complex neural network architectures. We also evaluate our proposed approach on diabetic retinopathy diagnosis task and benchmark with the state-of-the-art Bayesian deep learning techniques. We demonstrate MOPED method enables scalable variational inference and provides reliable uncertainty quantification.
Acoustic scene classification (ASC) has been approached in the last years using deep learning techniques such as convolutional neural networks or recurrent neural networks. Many state-of-the-art solutions are based on image classification frameworks and, as such, a 2D representation of the audio signal is considered for training these networks. Finding the most suitable audio representation is still a research area of interest. In this paper, different log-Mel representations and combinations are analyzed. Experiments show that the best results are obtained using the harmonic and percussive components plus the difference between left and right stereo channels, (L-R). On the other hand, it is a common strategy to ensemble different models in order to increase the final accuracy. Even though averaging different model predictions is a common choice, an exhaustive analysis of different ensemble techniques has not been presented in ASC problems. In this paper, geometric and arithmetic mean plus the Ordered Weighted Averaging (OWA) operator are studied as aggregation operators for the output of the different models of the ensemble. Finally, the work carried out in this paper is highly oriented towards real-time implementations. In this context, as the number of applications for audio classification on edge devices is increasing exponentially, we also analyze different network depths and efficient solutions for aggregating ensemble predictions.
This paper presents a novel method which simultaneously learns the number of filters and network features repeatedly over multiple epochs. We propose a novel pruning loss to explicitly enforces the optimizer to focus on promising candidate filters while suppressing contributions of less relevant ones. In the meanwhile, we further propose to enforce the diversities between filters and this diversity-based regularization term improves the trade-off between model sizes and accuracies. It turns out the interplay between architecture and feature optimizations improves the final compressed models, and the proposed method is compared favorably to existing methods, in terms of both models sizes and accuracies for a wide range of applications including image classification, image compression and audio classification.
Modern deep neural networks are well known to be brittle in the face of unknown data instances and recognition of the latter remains a challenge. Although it is inevitable for continual-learning systems to encounter such unseen concepts, the corresponding literature appears to nonetheless focus primarily on alleviating catastrophic interference with learned representations. In this work, we introduce a probabilistic approach that connects these perspectives based on variational inference in a single deep autoencoder model. Specifically, we propose to bound the approximate posterior by fitting regions of high density on the basis of correctly classified data points. These bounds are shown to serve a dual purpose: unseen unknown out-of-distribution data can be distinguished from already trained known tasks towards robust application. Simultaneously, to retain already acquired knowledge, a generative replay process can be narrowed to strictly in-distribution samples, in order to significantly alleviate catastrophic interference.
This paper proposes a zero-shot learning approach for audio classification based on the textual information about class labels without any audio samples from target classes. We propose an audio classification system built on the bilinear model, which takes audio feature embeddings and semantic class label embeddings as input, and measures the compatibility between an audio feature embedding and a class label embedding. We use VGGish to extract audio feature embeddings from audio recordings. We treat textual labels as semantic side information of audio classes, and use Word2Vec to generate class label embeddings. Results on the ESC-50 dataset show that the proposed system can perform zero-shot audio classification with small training dataset. It can achieve accuracy (26 % on average) better than random guess (10 %) on each audio category. Particularly, it reaches up to 39.7 % for the category of natural audio classes.
Adversarial audio attacks can be considered as a small perturbation unperceptive to human ears that is intentionally added to the audio signal and causes a machine learning model to make mistakes. This poses a security concern about the safety of machine learning models since the adversarial attacks can fool such models toward the wrong predictions. In this paper we first review some strong adversarial attacks that may affect both audio signals and their 2D representations and evaluate the resiliency of the most common machine learning model, namely deep learning models and support vector machines (SVM) trained on 2D audio representations such as short time Fourier transform (STFT), discrete wavelet transform (DWT) and cross recurrent plot (CRP) against several state-of-the-art adversarial attacks. Next, we propose a novel approach based on pre-processed DWT representation of audio signals and SVM to secure audio systems against adversarial attacks. The proposed architecture has several preprocessing modules for generating and enhancing spectrograms including dimension reduction and smoothing. We extract features from small patches of the spectrograms using speeded up robust feature (SURF) algorithm which are further used to generate a codebook using the K-Means++ algorithm. Finally, codewords are used to train a SVM on the codebook of the SURF-generated vectors. All these steps yield to a novel approach for audio classification that provides a good trade-off between accuracy and resilience. Experimental results on three environmental sound datasets show the competitive performance of proposed approach compared to the deep neural networks both in terms of accuracy and robustness against strong adversarial attacks.
In this paper, we investigate how to learn rich and robust feature representations for audio classification from visual data and acoustic images, a novel audio data modality. Former models learn audio representations from raw signals or spectral data acquired by a single microphone, with remarkable results in classification and retrieval. However, such representations are not so robust towards variable environmental sound conditions. We tackle this drawback by exploiting a new multimodal labeled action recognition dataset acquired by a hybrid audio-visual sensor that provides RGB video, raw audio signals, and spatialized acoustic data, also known as acoustic images, where the visual and acoustic images are aligned in space and synchronized in time. Using this richer information, we train audio deep learning models in a teacher-student fashion. In particular, we distill knowledge into audio networks from both visual and acoustic image teachers. Our experiments suggest that the learned representations are more powerful and have better generalization capabilities than the features learned from models trained using just single-microphone audio data.
This study investigated the waveform representation for audio signal classification. Recently, many studies on audio waveform classification such as acoustic event detection and music genre classification have been published. Most studies on audio waveform classification have proposed the use of a deep learning (neural network) framework. Generally, a frequency analysis method such as Fourier transform is applied to extract the frequency or spectral information from the input audio waveform before inputting the raw audio waveform into the neural network. In contrast to these previous studies, in this paper, we propose a novel waveform representation method, in which audio waveforms are represented as a bit sequence, for audio classification. In our experiment, we compare the proposed bit representation waveform, which is directly given to a neural network, to other representations of audio waveforms such as a raw audio waveform and a power spectrum with two classification tasks: one is an acoustic event classification task and the other is a sound/music classification task. The experimental results showed that the bit representation waveform achieved the best classification performance for both the tasks.
Classification of audio samples is an important part of many auditory systems. Deep learning models based on the Convolutional and the Recurrent layers are state-of-the-art in many such tasks. In this paper, we approach audio classification tasks using capsule networks trained by recently proposed dynamic routing-by-agreement mechanism. We propose an architecture for capsule networks fit for audio classification tasks and study the impact of various parameters on classification accuracy. Further, we suggest modifications for regularization and multi-label classification. We also develop insights into the data using capsule outputs and show the utility of the learned network for transfer learning. We perform experiments on 7 datasets of different domains and sizes and show significant improvements in performance compared to strong baseline models. To the best of our knowledge, this is the first detailed study about the application of capsule networks in the audio domain.
With the recent advancements in AI, Intelligent Virtual Assistants (IVA) have become a ubiquitous part of every home. Going forward, we are witnessing a confluence of vision, speech and dialog system technologies that are enabling the IVAs to learn audio-visual groundings of utterances and have conversations with users about the objects, activities and events surrounding them. As a part of the 7th Dialog System Technology Challenges (DSTC7), for Audio Visual Scene-Aware Dialog (AVSD) track, We explore `topics' of the dialog as an important contextual feature into the architecture along with explorations around multimodal Attention. We also incorporate an end-to-end audio classification ConvNet, AclNet, into our models. We present detailed analysis of the experiments and show that some of our model variations outperform the baseline system presented for this task.
In recent decade, many state-of-the-art algorithms on image classification as well as audio classification have achieved noticeable successes with the development of deep convolutional neural network (CNN). However, most of the works only exploit single type of training data. In this paper, we present a study on classifying bird species by exploiting the combination of both visual (images) and audio (sounds) data using CNN, which has been sparsely treated so far. Specifically, we propose CNN-based multimodal learning models in three types of fusion strategies (early, middle, late) to settle the issues of combining training data cross domains. The advantage of our proposed method lies on the fact that We can utilize CNN not only to extract features from image and audio data (spectrogram) but also to combine the features across modalities. In the experiment, we train and evaluate the network structure on a comprehensive CUB-200-2011 standard data set combing our originally collected audio data set with respect to the data species. We observe that a model which utilizes the combination of both data outperforms models trained with only an either type of data. We also show that transfer learning can significantly increase the classification performance.
We propose an efficient end-to-end convolutional neural network architecture, AclNet, for audio classification. When trained with our data augmentation and regularization, we achieved state-of-the-art performance on the ESC-50 corpus with 85:65% accuracy. Our network allows configurations such that memory and compute requirements are drastically reduced, and a tradeoff analysis of accuracy and complexity is presented. The analysis shows high accuracy at significantly reduced computational complexity compared to existing solutions. For example, a configuration with only 155k parameters and 49:3 million multiply-adds per second is 81:75%, exceeding human accuracy of 81:3%. This improved efficiency can enable always-on inference in energy-efficient platforms.
In this paper, we propose a multi-level attention model to solve the weakly labelled audio classification problem. The objective of audio classification is to predict the presence or absence of audio events in an audio clip. Recently, Google published a large scale weakly labelled dataset called Audio Set, where each audio clip contains only the presence or absence of the audio events, without the onset and offset time of the audio events. Our multi-level attention model is an extension to the previously proposed single-level attention model. It consists of several attention modules applied on intermediate neural network layers. The output of these attention modules are concatenated to a vector followed by a multi-label classifier to make the final prediction of each class. Experiments shown that our model achieves a mean average precision (mAP) of 0.360, outperforms the state-of-the-art single-level attention model of 0.327 and Google baseline of 0.314.
Explainable Artificial Intelligence (XAI) is targeted at understanding how models perform feature selection and derive their classification decisions. This paper explores post-hoc explanations for deep neural networks in the audio domain. Notably, we present a novel Open Source audio dataset consisting of 30,000 audio samples of English spoken digits which we use for classification tasks on spoken digits and speakers' biological sex. We use the popular XAI technique Layer-wise Relevance Propagation (LRP) to identify relevant features for two neural network architectures that process either waveform or spectrogram representations of the data. Based on the relevance scores obtained from LRP, hypotheses about the neural networks' feature selection are derived and subsequently tested through systematic manipulations of the input data. Further, we take a step beyond visual explanations and introduce audible heatmaps. We demonstrate the superior interpretability of audible explanations over visual ones in a human user study.